CA2271918A1 - Transversely coupled fiber optic sensor for measuring and classifying contact and shape - Google Patents

Abstract

The device according to the present invention includes an optical sensing device including first and second lightguides positioned side by side , that have been looped or bent and are preferably covered as a unit with a layer of transmissive material, wherein the lightguides are adapted to detect or sense the presence, location and identity of external media.

Description

TRANSVERSELY COUPLED FIBER OPTIC SENSOR FOR MEASURING AND
CLASSIFYING CONTACT AND SHAPE
Field of Invention This invention relates to fiber optic devices, and in particular optically sensing and measuring contact and shape, and classifying their properties.
Background of the Invention Prior art optical contact and shape sensor methods include detecting contact pressures by means of frustration of internal reflection, which can take on many forms.
Prior art contact sensors have been used to produce pressure contact images by having an external object deform a solid sheet into contact with a clear optical layer, the surface of the layer being illuminated by light which it reflects internally onto a camera.
The contacting solid or intermediate sheet replaces a covering layer of air in a light guiding structure with the 2o surface thereon, so that internal reflection at the point of contact is eliminated. In all cases, solids have a higher index than air, so even clear solids will tend to frustrate the reflections at the point of contact. The images collected by the camera or other vision system form pressure or contact maps that can be used for tactile sensing or object recognition.
It is also known in the prior art that transmission of light through a fiber may be frustrated by micro bending, or overbending, whereby the fiber is bent with a curvature sufficient to cause some of the higher modes within the fiber to escape through the cladding because they strike the cladding at an angle exceeding the conditions for total internal reflection. An example of a microbending sensor is given in (Hastings, M.C. et al., "Evaluation of special communications 3o grade fibers in interferometric and microbend sensors for measurements with ambient temperature fluctuations," SPIE Vol. 1795, Fiber Optic and Laser Sensors X, pp. 227- 235, 1992.) Microbending sensors, like the one indicated above, often use a shaped platen to impress multiple bends in a single fiber, thereby increasing the change in light throughput when pressure is applied.
Sensors that detect or measure the extent of contact of a liquid are also frequently based on frustration of total internal reflection. Prior art liquid level sensors include single point detectors that can detect presence or absence of liquid at a discrete point, and continuous sensors that measure the spatially continuous height of a liquid. Examples of prior art optical liquid level sensors include US Patent 4,038,650 to Evans, US Patent 4,353,252 to Jeans, US
Patent 4,788,444 to Williams, US Patent 4,039,845 to Oberhansli, US Patent 4,311,048 to Merz, US Patent 4,745,293 to Christensen, US Patent 3,448,616 to Wostl, US
Patent 4,880,971 to Danisch, US Patent 5362971 McMahon, and WIPO 86/03832 to Bellhouse.
Both Evans '650 and Jeans '252 describe point detectors that rely on frustration. Williams '444 and Oberhansli '845 describe point detectors wherein a similar frustration occurs within the wall of a containment. In the case of Oberhansli, the containment is the optical probe; Williams transfers the measurement to the wall of a clear container holding the liquid.
Merz '048 uses a cylindrical rod with circumferential V grooves to obtain a quasi- spatially continuous measurement based on frustration at successive V grooves.
A similar approach is taken by Christensen '293, except that non-circular V's comprise a grating, with V spacing corresponding to wavelength of the light. Successive mode extinction is used by Wostl '616 to achieve spatially continuous measurement using a tapered optical rod.
Danisch'971 uses a multi-layered probe to achieve spatially continuous measurement independent of index of refraction of the liquid surrounding a smooth probe.
Bellhouse 86/03832 uses a loop of optical fiber directly immersed in a liquid, the internal reflections being partly fivstrated by presence of liquid where there would ordinarily be air. A loop or loops, small enough in radius to cause egress of light, are used to increase the dependence of internal reflection on external media. Loops of this sort are best achieved with plastic optical fiber, which can be formed into tight curves without cracking. Maximum frustration in looped 3o sensors of this sort is achieved by removing or disturbing the clear cladding layer on the fibers, such as by roughening with sandpaper.

A similar looped approach is taken by McMahon '971. McMahon quantifies the performance of loops of various diameters from 1/8 to over 3/8 inches in diameter. Water produces a loss of from 10 to 20 percent of the throughput in air, while gasoline produces a loss of approximately 45 percent.
Examples of prior art for fiber optic shape sensing is given in Danisch patents: US 5,321,257, US 5,633,494, WIPO 0,702,780, and PCT WO 98/41815; and publications: Danisch, L.A., "Laminated BEAM loops," SPIE Vol. 2839, Fiber Optic and Laser sensors XIV, 12 pp., 1996; Danisch, L.A., Englehart, K., and Trivett, T., "Spatially continuous six degree of 1 o freedom position and orientation sensor," Fiber Optic and Laser Sensors and Applications, SPIE Conf. 3541A, Boston, MA, 1998; and Danisch, L.A., "Smartmove Human Machine Interface," Volume 1 of report to Canadian Space Agency, Project 9F028-7-7153/01-SW, 102 pp. + appendix, Dec. 13, 1998. In this prior art, the sensors report single degree of freedom curvature or complete shape using bend sensors alone, or an array of bend and twist sensors, respectively.
In (Danisch, L.A., "Smartmove Human Machine Interface," Volume 1 of report to Canadian Space Agency, Project 9F028-7-7153/01-SW, 102 pp. + appendix, Dec. 13, 1998) sensors arrays are described which report complete three dimensional shape of an object by measuring 20 bend and twist along a continuous flexure. In Danisch, '257, '494, '780, and '815, sensors are described which report monotonic (all positive or negative) curvature as a single valued output signal. The described sensors either have short sensing lengths, thereby ensuring monotonicity, or are long but used only in situations where monotonic curvature predominates. The prior art does not describe sensors designed to resolve or discriminate shape on the basis of a single output number, other than to correlate a single number with a net change in angle over a long sensor length. This "averaging" is described in Danisch, '494, col. 19, lines 9-14:
"Nevertheless, spaced emission surfaces are still an advantage for many sensors, as they can be used to sense average curvature over a greater axial length of the fiber. This can eliminate or reduce undesirable effects from large local changes in curvature, for instance due to the 30 presence of a foreign body under the fiber." Similar "averaging" is used in various other commercial devices to measure net angular changes. Examples include Penny and Giles goniometers and Virtual Technologies Inc. instrumented gloves, both of which use resistive bend sensors. These sensors report net angular change at the ends of a long flexural sensor, without regard to the intervening shape of the sensor.
New shape sensing art introduced in the description of the present invention includes single and double fiber sensor structures capable of generating a distinctive, yet single-valued output which can be used to determine the class of shape applied. Classes of shape include curvature parameters such as monotonic, inflected, number of inflections, local magnitudes beyond high or low limits, spatial frequency content, and number of peaks of a given spatial frequency content. The new art is distinguished from the above "averaging" technique, because it does respond to the intervening shape.
Summary of the Invention This invention comprises sensors made from single and multiple lightguides used as modulators, wherein the intensity of light passing through the lightguide(s) is changed (modulated) by a) external influences that change the shape of the lightguide(s), or b) external influences that change the optical media surrounding lightguide(s) shaped to be sensitive to contact with gases, vapors, liquids, or solids; without necessarily further influencing the shape of the lightguides.
Modulation occurs when the external media are applied to curved lightguide(s) or when external influences cause the lightguides to curve or to experience a change in curvature.
3o When the lightguides are straight, they are not influenced by external media, so description of the invention requires a consideration of lightguides in non-straight shapes including loops and sinuations.

It can be advantageous to sense curvature or shape of a surface. This can be useful in sensing disturbances to civil structures, movement between parts, impacted shape of an automobile for purposes of deploying air bags, or as a means of inputting information to a computer. For air bag deployment, the deployment decision is made by an on board computer that must receive input on the type of shape impacting. For instance, a very sharp shape, as from a utility pole, which might cut through the metal without activating an accelerometer used to detect impact, is to be distinguished from a very broad shape that would be detected by the accelerometer.
Other shapes such as multiple poles, or inflected and non-inflected shapes must often also be classified as part of a deployment decision. It is particularly important to detect shape of impact at the side of the car, where there is little material between the occupants and the colliding object.
For low cost shape sensing, such as for air bag deployment, it is desirable to use the smallest number of sensors possible. This can be accomplished by classifying shapes with a small number of long, flexible sensors, each designed to detect a certain class or classes of shapes.
The sensors are attached, for instance, next to each other along part of the side of the car, such as along a horizontal door beam. It is desired to obtain single-valued outputs from each sensor that can be interpreted individually to determine the class or classes reported by each sensor, and in concert to resolve the class of shape impacting and its rate of penetration.
Particularly difficult classes to distinguish from each other are single and double sharp impacts (e.g. a utility pole vs. two small vertical pipes). A sensor that simply integrates the absolute value of curvature along its length will tend to report the double impact with the same output as a particularly severe single-object impact. This is undesirable, because the two events often require different deployment actions. Another difficult pair of cases to distinguish includes broad and sharp shapes, such as those resulting from a guard rail and a utility pole, respectively.
It is also advantageous to know the position and weight of seat occupants in vehicles, for purposes of automated airbag deployment. It is further advantageous for this application to classify occupants by "configuration," such as "occupant is in an infant seat with sharp edges"
or "occupant is in a class with normal males." Information on these parameters can be found by measuring curvature or aspects of curvature, over the seat area, or chosen portions of the seat area.
It can also be advantageous to measure the presence, absence, and nature of media contacting a sensor structure. The same sensors used to make shape measurements can be made sensitive to media contact, by forming them into particular shapes that cause some of the light traveling through them to interact with the surrounding media.
It has been determined that with looped plastic fibers 0.25 and 0.5 mm in diameter, if any clear covering layer is used to protect the loops from external media, the maximum loss of light throughput that can be expected in the presence of water is 33%, and in most cases it is in the 10-20% range. Typical coverings are clear epoxy or a sheet of curved material such as polyethylene or polyester. If an array of many looped sensors is built with such covering materials, small variations in the thickness and integrity of contact with the covering materials will cause the losses at each sensor to differ greatly, so that it is di~cult to process the signals from the loops without a calibration table. However, frequent re-calibration is necessary because minor contamination of the coverings by dirty films or particulates will change the signal levels even more than the change due to liquid replacing air as the external medium.
It is possible to build an array with collections of prisms or similar structures, but this increases its size, expense, and vulnerability to damage.
It can be advantageous to use an array of sensors to determine contact over a surface or along a line, at discrete points. Such arrays can be used to measure height or presence of a liquid or contact of a solid. Such an array will be most useful if each element of the array produces a large change in signal due to contact, and only a minor change due to presence of contamination such as by films of oil, dirt, dust, or chemicals. Changes on the order of 90%
due to contact will make it unnecessary to calibrate the array, the signals being essentially binary. Yet, even though contact will be determined by essentially binary information, readings of the low light throughput after contact can be used to infer the nature of the contact, such as whether it is from a liquid of high or low optical index of refraction.
If a sensor experiences only a minor change due to contact, such as 20%, this drop in signal can be mimicked by a contaminant that also changes the signal by 20%, which can occur easily. This is to be contrasted to a sensor with a throughput which drops to 10% in the presence of water and to 5% in the presence of oil. 20% contamination would change these values to 8% and 4%, but they would still be quite useable to report that a) contact had occurred and b) the type of substance in contact.
The present invention enables very high modulation of light throughput due to contact with liquids, resulting in readings that are minimally influenced by layers of contamination.
For most applications, whether for shape or contact sensing, it will be necessary to have the elements of the array protected from the surrounding media. This is especially true if plastic fibers are used for the array. For looped arrays, plastic fibers are the most useful, because they can retain a sharp bend without propagating micro-cracks and eventually failing from a condition known as "static fatigue" in glass fibers. However, plastic fibers are quite wlnerable to damage from chemicals, such as organic solvents, and require coverings, especially if they are abraded to expose the core and increase coupling to the external media.
Unfortunately, clear coverings on individual loops and many other optical structures tend to defeat the measurement of external media. Because the coverings are of high optical index of refraction, they will act as frustrators, so that contacting materials will impart little additional change in throughput. Coverings that do not contact the fibers closely will also defeat the purpose of the sensor, because internal reflections will be maintained due to air between the covering and the protective layer. The coupling region is preferably covered through using a structure constructed of a lenticular lens layer covering the loops, although other suitable materials may be used.
The choice of optical fibers as a sensing means imparts qualities of safety 'and freedom from electrical interference, both due to the absence of electrical conductors within the sensing probe. It also makes possible very small and flexible sensing structures. The present invention benefits from the use of loops or bends in optical fibers as optical sensors, yet overcomes the disadvantages of prior art sensing loops. A particular advantage is the virtually complete modulation of the coupled light by the presence of liquids and solids, or by induced curvature, even though the device includes a protective coating. This makes the device relatively insensitive to the presence of contaminants on its surface. Contaminants cannot penetrate _$_ through the coating to the fibers, and have minimal effect on the measured values.
As such, various embodiments of the invention described herein use optical fibers or other light guides to achieve:
a) Detection of a liquid or solid contacting a sensing element with one bend that includes a clear protective layer;
b) Discrimination of the type of liquid contacting the element;
c) Measurement of liquid height with multiple elements or multiple bends;
d) Measurement of curvature at a single bend;
e) Discrimination of single from multiple shapes when multiple bends are applied;
Accordingly, one embodiment of the present invention is to provide an optical sensing device comprising a pair of optical fibers where the fibers are positioned side by side, and covered, as a unit, by a layer of optically transparent material having a convex arcuate outer surface.
Another important aspect of the present invention is the exploitation of the shapes of paired loops or bends to impart an optimal shape to a clear protective coating that couples the loops or bends optically. The coating has a shape determined by the loops or bends and the flow characteristics of the coating material.
Another important aspect of the present invention is the exploitation of the shapes of fibers laid side by side to impart an optimal shape to a clear protective coating that couples the fibers optically when they are bent. The coating has a shape determined by the fibers and the flow characteristics of the coating material.
Further aspects of the present invention include providing a sensor means that:
a) can be formed into small individual point sensors or a thin, quasi-spatially-continuous array;
3o b) does not expose the fibers directly to the media to be measured;
c) is capable of discriminating water from hydrocarbons;
d) is capable of detecting and measuring extent of contact with solids;

e) can be used to measure curvature and classify multiple curvatures to classify imposed shape.
f) can have an optical output of zero when not curved.
g) can be manufactured and instrumented at low cost.
Accordingly, it is another aspect of the present invention to provide an optical sensing device adaptable to detect or sense the presence, location and identity of external media.
Accordingly, it is another aspect of the present invention to provide a means of forming a single point sensor or an array of these sensors from fibers that have been looped or bent and 1 o covered with a layer or layers of durable material, and yet provide maximum modulation (90%
or more) when exposed to contact by liquids and solids. High modulation is desirable in achieving a liquid sensor that is able to be used to discriminate between different substances, such as a liquid and a gas, or between different liquids, especially in the presence of contaminating materials.
Accordingly, it is another aspect of the present invention to form sensors between fibers of a fiber optic ribbon cable, at desired lateral and axial locations along the cable, so that pressure at the sensor locations will generate a signal due to curvature of the cable, associated with the location of the sensor.
Accordingly, it is a fi~rther aspect of the present invention to provide a sensing strip with one or more fibers or other lightguides, so that impressed shapes will generate signal values from the sensing strip indicative of the class of shape impressed, based on combinations of curvatures contained in the shape class.
It is still a further aspect to accomplish any of the above this without use of special pre-formed optical shapes such as prisms or grooves, but rather to rely on the natural shapes of the loops or bends, covered with materials that do not require special forming processes, yet have a shape that produces the desired optical response. This also eliminates the need to connect 3o fibers to optical elements such as lenses or prisms, thereby simplifying the construction process and increasing the reliability of the instrument.

It is still a fi~rther aspect of the present invention to be able to construct a sensor array without requiring special alignment fixtures or molds to hold the fibers in place during and after construction.
It is still a further aspect of the invention to form flat sensor strips, that can be applied to flat or gently curved surfaces and used to sense or classify shapes when impressed upon the surface, or strips that can be curved at desired locations to activate a contact sensing capability at said locations.
It is still another aspect of the present invention to provide an inexpensive optical sensor device capable for use as a mufti-use sensor, such as a single sensor construction that can be used to sense liquid level, pressure, and shape.
According to one embodiment of the present invention, there is provided an optical sensing device comprising a first optical lightguide; a second optical lightguide with a portion of its length parallel to and in close proximity to the first lightguide to form a coupling region; where the first and second lightguides being covered within the coupling region; and wherein, light is coupled from the first lightguide to the second lightguide, when the lightguides are curved out of their plane within the coupling region. Desirably, the above cover for the coupling region is formed by a lens layer of optically transmissive material having a convex, arcuate outer surface.
In another aspect according to the above, the lightguides are mounted in a surface to be deformed by imposed pressures or shapes for purposes of determining classes and growth of impacted shapes.
According to a further embodiment, the first and second lightguides are modified for enhanced coupling, by abrasion, chemical treatment, heat forming, or notching, to lose and collect light in adjacent surface areas facing away from the plane of the lightguides in the coupling regions.
Desirably, the lens layer is formed on only one side of the plane of the lightguides, thereby enabling coupling only for curvatures of the lightguides which impose convex curvature on the lens layer.
According to any of the above embodiments, including a means for injecting light into the first lightguide, and means for detecting the intensity of light coupled into the second lightguide.
Further, according to the above including a means for injecting light into the first lightguide, means for detecting the intensity of light coupled into the second lightguide, and means for detecting the intensity of light carried through the first lightguide.
According to another alternative embodiment, the lens layer is formed on only the side of the plane of the lightguides containing the loss and collection areas, thereby enabling coupling only for curvatures of the lightguides which impose convex curvature on the lens layer.
Preferably, according any of the above embodiments, the transparent material comprises a synthetic resin., a heat dissolvable material or a chemically removable material.
According to another aspect of the present invention, the first and second lightguides are formed into curves out of the plane of the lightguides, within the coupling region.
In another alternative embodiment, there is provided a pressure or shape measuring and classifying sensor as described above, wherein the first and second lightguides are mounted on a surface to be deformed by imposed pressures or shapes.
According to another aspect of the present invention there is provided a pressure or shape measuring and classifying sensor wherein the lightguide diameter and coupling enhancement means are chosen to produce coupled light intensity that is maximal for only a single inflected shape and is attenuated when more than a single inflected shape is imposed.
Desirably, the lightguide diameter and coupling enhancement means are chosen to produce coupled light intensity that is maximal for shapes with large curvatures and minimal for shapes with minimal curvature.
According to another embodiment of the above pressure or shape measuring and classifying sensor, the lightguide diameter and coupling enhancement means are chosen to produce coupled light intensity that decreases for noninflected shapes and increases for inflected shapes, or to produce coupled light intensity that is minimal for inflected shapes and maximal for noninflected shapes.
According to various alternative embodiments described above, there is provided a pressure or shape classifying sensor comprising a first or second lightguide wherein the intensity of light that has passed through the lightguide is measured to classify the shape imposed on said lightguide according to the number of inflected curves, polarity of curvature, and magnitude of curvature.
In a further alternative embodiment according to the present invention there is provided a pressure or shape measuring and classifying sensor comprising a first plurality of sensors as described in an above alternative embodiment, exposed to a distribution of curvature within an extent; and a second plurality of sensors as described in an above alternative embodiment, exposed to a distribution of curvature within the extent; wherein the measurements of a pressure or shape distribution by the sensors are analyzed singly and in combination to classify the distribution of curvature within the extent according to absolute value, polarity, number of inflections, number of peaks, spatial frequency content, and location within the extent, and to 2o measure the time progress of the classifications. Desirably, the above may be used for determining classes and growth of impacted shapes in vehicles for purposes of airbag deployment. Preferably, the above sensor may be used for determining occupant position and weight in vehicles for purposes of airbag deployment.
According to another alternative embodiment of the present invention, there is provided a pressure or shape sensing array comprising sensors with coupling regions as described in any of the above embodiments, distributed over an area within which pressure or shape is to be measured at locations, wherein the sensor coupling regions are located to respond uniquely to pressure or shape at said locations and wherein the overall pressure or shape is inferred from 3o the individual sensor measurements. Preferably, the above sensor array comprises electrical conductors instead of lightguides, the coupling regions comprise electric coupling regions wherein coupling is modulated by bending, and the bending is determined by measuring electric current or voltage resulting from the coupling. Desirably, the above sensors are formed from adjacent fiber pairs of a fiber optic ribbon cable, wherein each coupling region occupies a known location along the axial extent of the cable.
According to any of the above embodiments, the sensor is preferably located between first and second mechanical layers, said mechanical layers containing structures capable of bending the sensors when pressure is applied.
According to any of the above alternative embodiments, there is provided a liquid or solid contact measurement sensor comprising wherein the coupling regions preformed into curves that couple light maximally when surrounded by a medium of low index of refraction and which couple light minimally when surrounded by a medium of high index of refraction.
Desirably, the above sensor is one in which a flexible surrounding material containing air at atmospheric pressure within is deflected by pressure from a liquid or solid medium without, to touch the curved coupling regions and produce changes in the measured intensity of light indicative of contact.
According to another embodiment, there is provided a liquid or solid contact measurement sensor comprising a sensor with coupling regions preformed into curves along its extent, each 2o curve of which couples light maximally when surrounded by a medium of low index of refraction and which couples light minimally when surrounded by a medium of high index of refraction.
According to any of the above alternative embodiments, the sensors may include a planar support member having an edge, and where the coupling regions may extend over the edge or may be spaced apart along and extend over the edge.
A liquid or solid contact measurement device as claimed in claim 1, said coupling region preformed into a curve with its apex exposed at the end of a tube covering the device.
3o Desirably, the above liquid contact measurement devices the intensity of coupled light when the device is immersed in liquid indicating the index of refraction of the liquid or level and composition of layered liquids.

According to another embodiment, there is provided a liquid or solid contact measurement sensor as described above including an array with spaced sensors, and motive means for changing the liquid or solid level with respect to the sensor array by a known displacement up to one intersensor spacing, the array measurement and said displacement being used to determine liquid or solid height or composition along a continuum.
In an alternative embodiment, there is provided a method of sensing a pressure or shape comprising the steps of providing a first optical lightguide, providing a second optical lightguide with a portion of its length parallel to and in close proximity to the first lightguide within a coupling region, covering the first and second lightguides within the coupling region, as a unit, by a lens layer of optically transmissive material having a convex, arcuate outer surface, transmitting light from a light source through the first optical lightguide, measuring the intensity of light coupled to 'the second lightguide through the lens layer, by measuring its intensity at the end of the second lightguide toward which said coupled light is directed, as a means of measuring curvatures within the coupling region.
In another alternative embodiment there is provided a method of sensing liquid or solid contact comprising the steps of providing a first optical lightguide, providing a second optical lightguide with a portion of its length parallel to and in close proximity to the first lightguide 2o within a coupling region, covering the first and second lightguides within the coupling region, as a unit, by a lens layer of optically transmissive material having a convex, arcuate outer surface, forming the coupling region into at least a single curve, transmitting light from a light source through the first optical lightguide, measuring the intensity of light coupled to the second lightguide through the lens layer, by measuring its intensity at the end of the second lightguide toward which the coupled light is directed, as a means of measuring the contact of liquid or solid and the index of refraction of the liquid or solid.
In a further alternative embodiment, there is provided an optical sensing device comprising an optical lightguide, an actuation operably device associated with the optical lightguide, 30 wherein the optical lightguide when deformed forms a coupling region adapted to transmit light along its length when the lightguide is curved out of its plane to the actuation device.
Desirably, the above device includes means for injecting light into the lightguide, and means for detecting the intensity of light coupled into the lightguide. Further, the above device is preferably provided with a cover for the coupling region formed by a lens layer of optically transmissive material having a convex, arcuate outer surface.
According to another aspect of the present invention, the above lightguide may be formed into curves out of the plane of the lightguide, within the coupling region.
Desirably, the above lightguide may be mounted on a surface to be deformed by imposed pressures or shapes.
According to an alterative embodiment according to the above embodiment, the device is 1 o provided for determining classes and growth of impacted shapes in vehicles for purposes of actuating an airbag actuation device. Desirably, the device further provides for determining occupant position and weight in vehicles for purposes of airbag deployment According to another alternative embodiment, the device may be used for determining classes and growth of impacted shapes along gaskets. Preferably, the above lightguide is mounted in a surface to be deformed by imposed pressures or shapes for purposes of determining classes and growth of impacted shapes.

The invention will now be described, with reference to the drawings.
FIGURE 1 is a side view of an optical fiber curved sufficiently to emit light, emitting light.
FIGURE 2 is an edge view of the same fiber in Figure 1.
FIGURE 3 is a side view of the fiber of Figure 1, receiving light.
3o FIGURE 4 is an edge view of the fiber of figure 3, receiving light.
FIGURE S is a graph plotting the optical throughput of a fiber as in figures 1 - 4, when bent at various curvatures.
FIGURE 6 is a side view of two fibers coupled by a lenticular layer along a coupling region, including a light source and receiver.
FIGURE 7 is a cross section through the two fibers on the line A-A of figure 6.
FIGURE 8 is an edge view of the fibers in figure 6.
FIGURE 9 is an edge view of the fibers as in Figure 6, bent within the coupling region.
FIGURE 10 is a perspective transparent view of curved fibers as in figure 9, including a lenticular layer along a coupling region.
FIGURE 11 is an edge view of the fibers of figure 10, showing light emitted from one fiber, reflecting internally within the lenticular layer, and entering the second fiber.
FIGURE 12 is a cross section on the line B-B of Figure 10, showing the coupling of light from fiber to fiber within the lenticular layer.
FIGURE 13 is a graph of the optical intensity throughput for light injected into one fiber of a coupled pair as in figure 10, for varying amounts of curvature of the structure.
FIGURE 14 is a perspective view of two fibers coupled by a lenticular layer, treated to have enhanced coupling areas along their upper surfaces, to enhance their ability to couple light from one fiber to another at lower curvatures than untreated fibers.
FIGURE 15 is a cross section of the line C-C of Figure 14.
3o FIGURE 16 is a perspective view of a coupled fiber structure as in Figure 15, but with multiple enhanced coupling areas applied along the upper surface, for piecewise continuous coupling between the two fibers.

FIGURE 17 is a schematic view representing the fibers as in Figure 16, with a longer coupling area indicated by the region containing overlapped lines.
FIGURE 18 is a schematic view representing the same two fibers of figure 17, bent in a single inflected shape within the coupling area.
FIGURE 19 is a schematic view representing the same two fibers of figure 17, but with two inflected shapes applied within the coupling area.
FIGURE 20 is a graph showing total throughput vs. number of separate curves, for a mathematical model of a coupled fiber structure as in Figure 17, for different attenuations at each curve, and a normalized throughput of 1.0 for each curve, before attenuation is applied to each coupled throughput.
FIGURE 21 is a schematic view of a coupled fiber sensor as in Figure 17, with multiple sinuations that can be used to sense the level of a liquid or amount of contact with a solid surface. The fibers include a turnaround loop so that light source and detector may be co-located.
2o FIGURE 22 is a schematic view of a single fiber sensor with a light source at one end and a light intensity detector at the other end.
FIGURE 23 is a schematic view of a fiber sensor with two parallel runs of fiber coupled by a loop, so that source and detector are co-located and the net throughput is a product of the throughput of individual purposely imposed enhanced coupling areas along the fibers and curvatures imposed on the fibers by an external force, said enhanced coupling areas optionally having different characteristics on each fiber.
FIGURE 24 is a schematic view of a coupled fiber structure with one fiber extended, so that 30 throughputs may be measured for light that traverses one fiber from beginning to end, and for light that traverses in lenticularly coupled fashion from one source on one fiber to a detector on the other fiber.

Figure 25 is a schematic view of a coupled fiber structure of Figure 24, with loops incorporated so that all sources and detectors may be co-located and more than one fiber run traverses the sensor area, each run being coupled by a loop to the next run, so that detected signals are the product of multiple runs.
FIGURE 26 is a perspective view of a lenticularly coupled sensor in the end of tubing.
FIGURE 27 is an edge view of the sensor of Figure 26, with the transmitting fiber on the left.
1 o FIGURE 28 is an edge view of the sensor of Figure 27, with the transmitting fiber on the right.
FIGURE 29 is a perspective view of two elements of an array of lenticularly coupled sensors, built on the edge of a thin band of steel.
FIGURE 30 is a simplified view of a complete array of four paired loops or bends attached to an interface box.
FIGURE 31 is a simplified view of an alternative embodiment of the present invention 20 including a complete array of four paired loops attached to an interface box, arranged in a standpipe to measure liquid height.
FIGURE 32 is a pressure sensor array formed from fiber optic ribbon cable, with coupling zones formed at discrete locations between fiber pairs by forming a clear lenticular structure at each location.
Having thus generally described the invention, reference will now be made to the accompanying drawings illustrating preferred embodiments.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
As illustrated in Figures 1 and 2, a fiber 10 curved sufficiently, will lose light along its periphery because some of the higher modes are unable to reflect internally within the fiber core. For simplicity, a thin cladding layer of low index of refraction, typically surrounding the core of all step index fibers, is not shown in the figures unless otherwise indicated. The cladding 24 layer is used to maintain internal reflection within a straight fiber even if comes into contact with solids or liquids. If there is no such contact, air will serve as a low index of 1 o refraction material to maintain internal reflections. Whether or not there is a cladding layer, a fiber curved below its minimum bend radius will begin to lose light. Example rays of light YY
are shown leaving the fiber 10 where it is curved.
Although the term "fiber" for "fiber optic" is used throughout the description, it is meant to apply generically to lightguides of various types, including clad or unclad bars of clear plastic or glass, of round; rectangular, or other cross section, capable of guiding light within the bar due to internal reflection.
Figure 2 shows the fiber 10 of Figure 1 from the edge. Example rays of light YY are seen to 2o be emitted not only within the plane of the loop, but within a cone of angles bisected by the plane of the loop.
As shown in Figure 3, a fiber 10 bent suflzciently to lose light is also capable of receiving light within the same range of angles that it can be emitted. Example rays YY are shown entering the fiber 10 where it is curved.
Figure 4 shows the fiber 10 of Figure 3 from the edge. Example rays of light YY are seen to be received not only within the plane of the loop, but within a cone of angles bisected by the plane of the loop. After entering the receiving fiber, the light will continue to propagate down 30 the fiber 10 within the cone of angles determined by the index of refraction of the core and that of its cladding 24, or other surrounding medium.

Figure 5 is a curve illustrating the loss of light throughput for a 0.25 mm diameter multimode step index plastic optical fiber 10 bent in a 180 degree circular curve (a "C"
shape) at various radii in air. Throughout this description, throughput is defined as intensity light collected at the output of an optical system under conditions of constant light input. In this figure, unattenuated throughput is represented by a value of 1Ø The fiber 10 begins losing light at a radius of approximately 5 mm, and continues to lose more and more light as the radius decreases. Larger fibers begin to lose light at larger radii. For instance, a 0.5 mm fiber will begin to lose light at a 10 mm radius, a 1.0 mm fiber at 20 mm, etc. A similar graph could be generated showing the receptivity of a fiber to external light along a curved section, with 1 o receptivity to a wider range of angles corresponding to smaller radii of curvature.
Figure 6 illustrates a dual fiber 12, 14 sensor structure with a source 16 and receiver 18 of light, each coupled to a different straight optical fiber. The fibers are parallel and in close proximity within a "sensing zone" 20 that is covered by a lenticular layer 22 (Figure 7) of clear material. Normally, no cladding is removed from the fibers, although other forms are possible using lightguides having no cladding, such as acrylic rods. The lenticular layer 22 may be applied to the full length of the overlapped fibers 12 and 14, or to a subset of the length.
For convenience or to provide mechanical protection, the same material may coat the non-overlapped fibers.
Figure 7 is a cross section through the overlapped portion of the sensor structure within the sensing zone 20. By way of example, in this figure, the cladding 24 on the fibers 12, 14 is shown. The lenticular layer 22 is shown as an oval shape surrounding the fibers. The following conditions apply to the lenticular layer:
a) It can be of any shape capable of reflecting light from one fiber to the other along lines defined by the emission and receiving characteristics of fibers curved below their minimum radii of curvature. Typical shapes are convex as shown or flat. Even concave shapes will serve.
b) It can be on one side of the fibers or both. If it is on one side, the sensor will only function when it is curved convexly on that side.

c) It can be of any index of refraction higher than that of air. Preferably, it will have an index in the 1.5 or higher range, typical of most materials capable of coating fibers.
However, the sensor will function when lower index materials are used, such as silicones in the 1.4-1.5 range of index of refraction.
d) It should have sufficient clarity to transmit light the short distance from one fiber to the other, but need not be of exceptional clarity. Ordinary epoxies, urethanes, casein resins and other coatings will fi~nction well for the short path encountered by light traveling from one fiber to the other.
e) It may be covered by a layer of other material of low index of refraction, for protective purposes or for cases where the sensor is to be used primarily as a means of sensing shape, with minimum coupling desired to external media.
The fibers within a lenticular layer 22 should be in close proximity, or the lenticular layer 22 must be made thicker to enable light to couple at the necessary angles for propagation when the fibers are bent. In most embodiments, the fibers will be touching in the overlapped regions, but can be several millimeters apart if the lenticular layer 22 is deeper.
2o Preferred methods for forming a lenticular layer 22 on adjacent parallel fibers include:
a) Spreading a synthetic light transmitting resin on the fibers with a spatula to fill the grooved space between the fibers. The resin will take on a flat or crowned cross section due to surface tension effects while curing.
b) Applying a continuous bead of synthetic light transmitting resin on the fibers through a syringe tip, with size of the bead controlled so that when it is curing, it will fill the grooved space between the fibers, taking on a crowned shape in cross section.
30 c) Applying an excess of synthetic light transmitting resin on the fibers, and wiping it offwith gloved fibers or flexible spatulas so that when it is curing, it will fill the grooved space between the fibers, taking on a crowned shape in cross section.

d) Any of the above, where the grooved space on both sides is covered with resin simultaneously, the fibers being suspended in air.
e) Any of the above, where the fibers are held together with temporary adhesive while the resin is being applied.
Figure 8 shows the sensor structure of Figures 6 and 7, in an edge view. When it is straight, there is negligible coupling of light from one fiber to the other, so there is no signal detected by the light receiver. In the present invention, evanescent coupling between adjacent fibers, 1 o which arises from photons having no finite boundaries according to the wave theory of light, is not considered to be a significant contributor to any of the coupled light.
Figure 9 shows the sensor structure of Figures 6 and 7, again in edge view, but with the sensing zone 20 bent in a 180 degree curve at a curvature where the fibers are capable of emitting and receiving light.
Figure 10 shows the sensor structure of Figure 9 in more detail, including the two curved portions of the fiber in close proximity to each other, and surrounded by the lenticular layer 22.
Figure 11 shows the sensor structure of Figure 10 in edge view, including arrows indicating the paths followed by example rays leaving the emitting fiber, reflecting internally within the lenticular layer 22, and entering the receiving fiber.
Figure 12 shows the sensor structure of Figures 10 and 11 in cross section, including arrows indicating the paths followed by example rays leaving the emitting fiber, reflecting internally within the lenticular layer 22, and entering the receiving fiber. In this figure the cladding layers are shown on the fibers.
3o Figure 13 is a graph of the coupled throughput of the sensor structure of Figure 9, for a 180 degree, or "C" shaped curve, for different radii of the curve. For radii above the minimum bend radius of 5 mm, there is very little or zero throughput, whereas the throughput rises dramatically as the curvature (curvature is the inverse of radius of curvature) increases.
Figures 14 and 15 illustrates a sensor structure including overlapping fibers surrounded by a lenticular layer 30, in a sensing zone 32. In this case, the fibers are treated to lose light along narrow strips by abrading or other methods described in patents by Danisch, US
5,321,257, US 5,633,494, and WIPO 0,702,780 to increase their ability to modulate throughput in response to bending. These treatments also serve to couple light into the fiber, so serve as a means of "enhanced coupling" between the fibers even when they are straight or bent without violating the minimum bend radius. In this alternative embodiment, the treatment is applied to 1 o enable use of the sensing zone 32 at large radii of curvature. For instance, the structure used to generate the graph of Figure 13 has no throughput for radii above approximately 5 mm, but the same configuration with abraded zones as in the present figure begins to lose light for radii of 10 cm, and wider or deeper "enhanced coupling" zones can be applied to achieve coupling even for straight fibers. If the enhanced coupling zones are restricted to one side of the fiber, coupling will increase for bends that curve the enhanced coupling zones convexly, even for large radius bends, but coupling will normally be minimal or zero for bends in the other direction until microbending effects begin to take place (e.g. 5 mm radius for the fibers in this example). An exception to this can be created with fibers that have enhanced coupling zones with very high loss. Then, coupling may be nonzero even for straight fibers, and will decrease 2o to zero when the enhanced coupling zones are increasingly concave. Given that loss strips may be applied in various lengths, spacings, and conformations given in Danisch'257, '494, and '780, including on one side or two sides, or circumferentially around the fibers, many throughput vs. curvature characteristics are possible. An example of bipolar enhanced coupling zones distributed along the fibers in quasi-continuous fashion is shown in Figure 16.
The distribution is made quasi-continuous to prevent depletion of loss modes as described in Danisch'257, '494 and '780. It is also desirable to displace the collection zones 40 "downstream" (away from the light source) of the loss zones 42, because light loss occurs at angles directed away from the source, and collection is optimal for angles directed toward the source. A typical downstream displacement for 0.25 mm fibers is 0.5 to 1 mm.
Characteristics of coupled fiber structures with different types of enhanced coupling zones include:

a) Monopolar: throughput responds equally to curvatures of either polarity.
This can be achieved with untreated fibers in the microbending region, or with fibers treated on both side or circumferentially within a larger class of radii. Monopolar sensors respond to the absolute value of curvature.
b) Bipolar: throughput increases for one polarity of curvature, decreases or is unchanged for the other. This can be achieved with fibers treated heavily on one side.
c) Nonuniform or nonlinear: throughput responds differently for different classes of 1 o curvature. For instance, the throughput of coupled fibers treated with enhanced coupling zones will respond with gradual changes to large radius bends, and will respond with increased sensitivity to bends within the range of radii where microbending effects predominate.
d) Spatially varied: If enhanced coupling zones are applied with varied spacing and length, coupled fiber structures may be achieved that have zones of sensitivity and insensitivity along their lengths. Curvatures applied to the fibers may tend to miss the sensitive zones, or be poorly sampled by the sensitive zones, or not have any effect or minimal effect in zones that are purposely designed to be insensitive or minimally 20 sensitive to curvature.
All of the light coupled from one curved fiber to another in a coupled fiber sensor as in Figure 9 relies on internal reflection within the lenticular layer 22. Without internal reflection, which relies on the presence of a medium of low index of refraction compared to that of the lenticular layer 22, surrounding the lenticular layer 22, coupling will be reduced to a lesser value or to zero, depending on the index of the surrounding medium and its extent of contact with the curved portion of the fibers. The following are examples of measured throughput for various media surrounding lenticularly coupled loops with a radius of l mm, referenced to a normalized value of 1.0 for air:
a) Air (index of 1.00): 1Ø
b) Water (index of 1.33): 0.08 c) Motor oil (index of 1.43): 0.04 Many other hydrocarbons have indices in the 1.4 to 1.5 range, and produce results similar to that of motor oil, and always easily distinguished from those of water or air.
When solids come in contact with a curved coupled fiber sensor, the coupled light is also frustrated, to a degree dictated by the contact surface area. The surface area of liquid contact also determines the throughput for a single loop, the throughput rising to a maximum for total contact.
Figure 17 shows a coupled fiber structure in schematic form, with the fibers 2, 14 drawn as two parallel lines, the overlapped portion 56 representing fibers coupled with a lenticular structure.
Figure 18 shows the fibers of Figure 17 with a curvature applied in the shape of an inflected curve 50. This may be thought of as the shape of a dent applied to the side of an automobile, or curves in a flexible coupled fiber pressure sensor caused by pressure from an object such as a finger, or could represent fibers that are held in constant curves so that liquid or solid contact may be sensed at the curves. The curves are said to be inflected because they include positive and negative values of curvature.
Figure 19 shows the fibers of Figure 17 with two inflected curves 52, 54 applied at different locations along the coupled portion of the fiber.
Figure 20 depicts the calculated throughput of a sensor as shown in Figure 19, but with the number of inflected curves varying from one to five. It will be described in more detail below as the equations for throughput are derived.
Figure 21 depicts a sensor as in Figure 17, but with multiple sinuations applied. It will be described in more detail below in the context of liquid sensing. It is presented here in association with Figure 19, as an example of a coupled fiber sensor with multiple curves. In the case of Figure 19, the multiple curves result from an imposed shape. In Figure 21 they are purposely imposed to enable contact sensing of a liquid.

We now address the issue of throughput for single and multiple curves along the fibers. First, we will consider a single fiber like one of the fibers shown in Figure 17, but with source and detector connected to the two ends. Such a fiber is shown in Figure 22 and again in Figure 23, where a loop 60 is incorporated so that source and detector may be co-located, and two runs of fiber coupled by the loop may be treated differently by means of purposely imposed enhanced coupling zones of different constructions, so that each run contributes in a different way to the net throughput signal when curves or contacts are imposed. Within a single fiber run or runs coupled by a loop of the same fiber, the net throughput is a product of the losses at each curve. If a fiber has n curves, each curve i in the fiber has a throughput Ei resulting from microbending or purposely imposed loss zones, and if Es is the net throughput of such a single fiber (normalized to unity for unattenuated throughput) , then (Eq. 1) Es = j~Ei, where the product ~ is taken from the initial through the nth curve i.
Thus, if 6 equal curves each of throughput 0.5 are applied, the net throughput Es=(Ei)°=O.Sx0.Sx0.5x0.Sx0.5x0.5=0.016.
Three curves of throughput 0.5 result in Es= 0.125, or 8 times more throughput than 6 curves.
The "6 curve" vs "3 curve" example is relevant to comparing the effects of a single dent to those of double dents, where each dent includes a positive and two negative curves of equal magnitude, and the coupling is monopolar (equal for any polarity of curvature). At first it might seem useful to have multiple dents produce less throughput than a single dent, but it must be remembered that in a single fiber sensor, reduction in throughput is a signal indicating increasing depth of a dent, so that multiple small dents will mimic a single large dent. This is the opposite effect to that desired for airbag deployment, for which one may wish to ignore multiple small dents. However, a single fiber sensor may be useful in detecting sharp vs. broad dents. The former will produce more attenuation than the latter, due to the larger curvatures implied by sharp vs. broad dents.
By varying the type and placement of loss zones, it is possible to create single fiber sensors that respond differently to different shapes. In the example above, the sensor responds with increasing attenuation to higher curvatures or more dents. If enhanced coupling zones of minor attenuation are applied to one side of the fiber, then a saturated response in throughput to bends of a given polarity can result. As shown in Danisch, US 5,633,494, and WIPO
0,702,780, a lightly treated (an imposed enhanced coupling zone with small attenuation) sensor fiber will exhibit a throughput that saturates at a high value for concave bends of the treated zone above a characteristic curvature, and which continues to decrease for increasing bends in the opposite, convex direction.
If such a sensor is used to detect dents, it will have an accentuated response to bends in the convex direction, so that it can be used, for instance, to classify inflected dents from noninflected (monotonic) dents. This is done by applying the sensor so that norunflected dents cause convex curvature of the treated zones, thereby causing an increase in throughput that saturates, whereas inflected dents will cause a large net decrease in throughput due to the imposition of two concave curvatures with unsaturated decreasing throughput at the edges of a single convex curvature that saturates at a low value of increasing throughput. If the sensor is inverted so that noninflected curves cause a decrease in throughput, they will be sensed as a non-saturated decrease in throughput. Inflected dents will also cause a decrease in throughput, representing the product of one convex curve (large unsaturated decrease) response with two concave curve responses (small saturated increases). Sharp inflected dents will produce larger drops in throughput than broad inflected dents, since most broad inflected dents will have two concave edges that are below the saturation limit, and will have a net zero product, the net result of one small decrease and two small increases.
Others may be made into sensors that have no treatment, in which case they will respond with attenuation that increases for sharper bends or more bends, without regard to polarity of curvature within the microbending range; or with loss zones on both sides, so that response also disregards polarity but is not restricted to the microbending range; or with loss zones on one side but such that response is bipolar (regards polarity of curvature) over a broad range of curvatures (as opposed to the bipolar saturated response described above).
Treated fibers with a bipolar response may also be used to classify noninflected shapes from inflected shapes.
Inflected shapes that begin and end with zero curvature have a net curvature of zero regardless of the curvatures within the shape, so will be "invisible" to a bipolar sensor. In contrast, a bipolar sensor will detect noninflected shapes easily, as they have a net positive or negative curvature.
Multiple sensors with different characteristics may be added to a door panel or the side of a vehicle to classify impacts by shape and to deploy air bags depending on the shape class and the magnitude of the shapes over time. The outputs of the sensors may be combined arithmetically in an electronic processor by conventional analog or digital means.
Combinations include arithmetic addition or subtraction, or logical AND and OR
operations, based on each sensor triggering a binary logic state indicating the class of impact shape detected, and these logic states then being resolved by AND and OR
combinatorial logic. It is also possible to combine responses within a single fiber, by providing multiple runs of the fiber across the region to be sensed, each coupled to the next through a turnaround loop. The combination will be a product of the individual sensor characteristics, which may be varied by type of treatment and by not inverting or inverting the treated portions with respect to convex or concave shapes.
A sensor in which there is coupling between the two fibers due to a lenticular structure has throughput characteristics that are related to the attenuation effects exhibited by a single fiber, but modified by coupling effects that tend to counteract the attenuations of a single fiber. This suggests combining single fiber sensors with lenticularly coupled fiber sensors to better classify shapes. Such a combination will be described below, but first we will describe a lenticularly coupled sensor. For such a sensor to have a nonzero output, the curves must be of sufficient curvature to cause coupling from one fiber to the other, due to microbending or purposely applied enhanced coupling zones. For each curved region, there will be light coupled within the curved region. However, a curved region also attenuates light passing through either fiber toward other locations along the sensor structure, both in the emitting and receiving fiber. This causes the amount of light reaching the detector from multiple curved zones to be equal to or less than the amount from a single curved zone, for curvatures that each attenuate the light passing by them by half or more. For smaller attenuations, the throughput may increase and then fall offwith increasing numbers of curves, or even continue increasing as more and more curves are added.

The attenuation of signals from multiple curves is explained in the following way: Each curved zone attenuates light passing through it, from any source, due to microbending or purposely formed enhanced coupling zones or regions. Light coupled across at any curved zone will encounter transmission fiber losses from curves between the light source and the coupling zone or region, and receiving fiber losses from curves between the coupling zone or region and the detector. The number of curves imposing losses will be the same for any coupling zone, since zones nearer the detector will have fewer receiver fiber losses but more transmission fiber losses, and zones nearer the source will have fewer transmission fiber losses but more receiver fiber losses. For instance, a fiber structure with 6 equally curved zones 1 o along its length, wherein each curve imposes a local drop in throughput from 1.0 to 0.5, will have an overall throughput of 0.5 x 0.5 x 0.5 x 0.5 x 0.5 x 0.5=.016 for each zone of coupling.
We apply only one loss figure at the region of coupling because coupling is distributed across the length of the curvature. If we consider that each curve would couple a unit amount of light if no attenuations occurred, then the total coupled throughput will be the sum of the attenuated unit amounts. Each attenuated coupled amount is 1.0 x 0.016 =
0.016, so the sum, or total throughput is .016 x 6 = 0.096. In contrast, an individual curve would have a throughput of 0.5 x 1.0=0.5, or more than 5 times more than the combined signal from the 6 curves. In this example, three curves would produce a throughput signal of 0.125 x 3 = 0.38 compared to the signal of 0.096 for six curves. The "6 curve" vs "3 curve"
example is relevant 2o to comparing the effects of a single dent to those of double dents, where each dent includes a positive and two negative curves of equal magnitude, and the coupling is monopolar (equal for any polarity of curvature). The double dents produce a signal almost 4 times smaller than a single dent of the same magnitude.
If curves of different magnitudes are applied at different locations along a coupled fiber structure, then the net effect is of course complicated by the different amounts of coupled light, and the differing attenuations, at each curve. However, increasing curvature is associated with increasing coupling, so increased coupling is generally accompanied by increased attenuation, so the net throughput from multiple curves changes only slightly with changes in 3o curvature of the individual curves. This is to be contrasted with the very robust increase in throughput for a single curve with increasing curvature.

Even when the curvatures are varied along the coupled fibers, the attenuations are the same for light coupled at any of the curves, because both transmitter and receiver fibers have the same curves.
The net throughput, Ec, for a coupled fiber sensor with n individual curves, each with unattenuated coupling Ki for light between the fibers, and attenuation Ai for light traveling down a fiber, is given by:
(eq. 2) Ec = QIAi)(~Ki), where the product ~ and sum ~ are taken over all i from initial to nth.
Figure 20 is a graph of Ec vs. n, for various Ai from 0.1 to 0.9. The graph was created from a mathematical model in which each coupled throughput is assigned a normalized value of 1.0 before attenuations are applied. It can be seen that for increasing n, the net throughput either decreases, rises and then decreases, or continues to increase, as Ai is varied. For most fibers, the higher Ai values apply unless the curvatures are at the low end of values that produce measurable coupling. Note that in an actual sensor the unattenuated coupled throughputs would not have values of 1Ø Instead, they would be small for low curvatures and large for high curvatures, so that the absolute magnitudes of the family of plots in the graph would be different from that shown, usually opposite to that shown. However, the graph is only meant to indicate the variations for each single plot of the family, as the number of imposed curves is varied (by "plot" within a family, we mean what is usually called a "curve"
within a family of curves, but have avoided the conventional term to avoid confusion with the spatial curves applied to fibers).
The net throughput can be used as a measure of the shape of indentations in the side of a vehicle, such as to emphasize airbag deployment for sharp single dents vs.
sharp double dents.
By varying the type and placement of enhanced coupling zones and Ai, it is possible to create coupled fiber sensors that respond differently to different shapes. For instance a sensor with large fibers or with added enhanced coupling zones will respond to broader shapes than a sensor with very small fibers and/or no purposely added enhanced coupling zones. As another example, a sensor with low attenuation per bend might have an output that increases according to the number of dents, while another with high attenuation might have an output that decreases with the number of dents. As a further example, a coupled fiber sensor with added enhanced coupling zones could respond to a broad monotonic curvature, while an untreated sensor would not respond at all to a broad monotonic curvature.
Multiple sensors with different characteristics may be added to a door panel or the side of a vehicle to classify impacts by shape and to deploy air bags depending on the shape class and the magnitude of the shapes over time.
The single fiber sensors of Figure 22 may be combined with lenticularly coupled fiber sensors of Figure 23 by extending the transmission fiber of a lenticularly coupled sensors and fitting it with a second detector 18b as in Figure 24. The second detector 18b will respond to the net throughput of the transmission fiber 12 alone, the first detector 18a responding only to light coupled from one fiber to the other. The outputs of the two sensors may be combined arithmetically or logically as described for combination of the individual types of sensors.
The single fiber sensors of Figure 22 or Figure 23 are able to broaden the classification abilities achievable with a double fiber coupled sensor, since they have a different throughput equation when in use. In the latter embodiment, it is the same equation whether or not the "single fiber sensor" is a stand-alone device or is a portion of a double fiber sensor. While a double fiber coupled sensor has a throughput equation including product and sum terms (see eq. 2 above), the equation for a single fiber sensor has only the product term. Shapes including multiple curves, particularly multiple inflected curves, will result in different outputs from the single and double fiber sensors, beyond the obvious distinction that single fibers have a throughput even when straight. An important property of a single fiber sensor with a nonlinear response to curvature magnitude is that its output can be made to decrease as more peaks or inflections are imposed, whereas a double fiber sensor can respond to the addition of peaks with little change in output. This means that it is possible to use the single and double fiber sensors together to resolve the number of peaks, and also to gain information about the magnitude of the applied curvatures. For example, in the case of a single fiber sensor that saturates at a certain positive curvature, its throughput remains constant for curvatures above a certain positive value. For a single dent (an inflected curve, or "peak") containing a curvature beyond the saturation value, its throughput will decrease. As more dents are applied, its net throughput will continue to decrease. In contrast, the throughput of a double fiber coupled sensor can be made to remain approximately the same as more dents are imposed, due to the summing term of equation 2. By comparing the outputs of the single and double fiber sensors, one may obtain information about the number of dents and their overall magnitude. The single fiber output will indicate number of dents, whereas the double fiber output will indicate magnitude of the curvature of the "sharpest" (highest spatial frequency) dent.
The ability to classify shapes is a feature of sensors that are long compared to the highest 1 o spatial wavelengths present in the shape, particularly if it is possible to introduce nonlinearities or other local modifications to the magnitude response of the sensor.
Nonlinearities permit obtaining useful outputs from sets of curvatures that would otherwise sum to zero. Other "tuning" factors are given in the example introduced in the following paragraph. The ability to tune the response of a long sensor to sense shape is an important aspect of the present invention, which distinguishes it from prior art.
Shape classification may be performed with sensors of any technology that are able to combine curvature information along their lengths. Examples include capacitive bend sensors and resistive polymer bend sensors, as well as strain gauge bend sensors. An example of 20 classification of shapes is given below, applicable to any sensor technology capable of measuring the integral or product of curvature along its length.
Definitions used in the example are:
Curvature: C= dA/ds, where 8 is the angular orientation of a space curve and s is the distance along that curve, regardless of its shape. The space curve is taken to be in Cartesian space, with x and y coordinates. Curvature devolves to d6/dx for shallow shapes, which is similar to the derivative of slope, or (dy/dx)'. This approximation is used in deriving beam equations (ref. Crandall and Dahl, An introduction to the 3o mechanics of solids, McGraw-Hill, NY, p. 362, 1959), but does not hold well for sharp dents.

Monotonic shape: a shape without inflections of curvature, i.e. the curvature is all of one sign (positive or negative).
Inflected shape: a shape that contains both positive and negative curvatures.
Note that this definition is for curvature (8,s realm). In the x,y realm a shape can have monotonic (all positive or all negative) slope (such as a circular dent) but will still be inflected in the curvature realm.
Dent: a synonym for inflected shape.
Peak: a synonym for inflected shape.
Integrated curvature: The integral of curvature along s. This is typically what a distributed fiber sensor reports, for all shapes applied along its length s.
Local integration: It is important to point out that a fiber (or many other distributed sensors) perform local integration, reporting a single number at the output.
This is what can make them "smart," if we are able to "tune" what is integrated. Local integration is what produces a zero result for a linear curvature sensor exposed to an inflected curve that starts and ends with zero slope. It is important to note that, for instance, the local integral of the absolute value of curvature will produce a large result, whereas the absolute value of the integral of curvature will produce a zero result for the inflected curve mentioned earlier in this paragraph. Sensors may integrate curvature along their lengths, or form a product of local or incremental curvatures. A
product, if treated logarithmically, becomes an integral, as the logarithm of a product is the sum of the logarithms of the product factors. Also, a product of large (close to a normalized value of 1.0) throughputs that decrement by a small amount behaves approximately as 1-(the sum of the decrements). Therefore, it is frequently permissible to view a product as an integral. Either can be used to perform classification according to the methods given in the description of the present invention.
Tuning: Selecting the local sensors along a fiber (or other sensor) so that the integral is taken over functions including the following functions or their combinations (others are possible as well):
a. absolute value of curvature.
b. positive curvature only.
c. negative curvature only.
d. curvature that saturates at a chosen positive magnitude.
e. curvature that saturates at a chosen negative magnitude.
f. curvature selected through a spatial comb filter.
g. responses that vary along s.
In the present example, we present a table containing important classes for discrimination of accident events. It shows discrimination is possible for most of these by using two sensors in combination. The sensors are of two types, called "1" and "2" with characteristics as described in the table. A third sensor described below is sufficient to discriminate a one remaining "problem" case. The third sensor is similar to sensors l and 2 but employs a specific saturation point for determining a particular class. The characteristics of sensors 1, 2, and 3 are included in the description of various single and coupled fiber sensors given earlier in the description of the present invention.
(Key to the table: si = sharp, inflected; sit = two si dents; bi = broad inflected; bm = broad monotonic, Ixx = output of sensor 1 for xx dent; 2xx = output of sensor 2 for xx dent; xx =
bi, bm, si, or bi).

Class Sen dent inflectionsno. output can aliasing resolve of sor width of of (pos pos alias is s as dent no. curvatureden is with resolved (C) is larger (NA--noby:
than t pos) aliased) si 1 sharp inflected1 pos lbi 2 si sit 1 sharp inflected2 pos pos NA sit bi 1 broad inflected1 pos 1 si 2 bi bm 1 broad monotonic1 neg NA bm si 2 sharp inflected1 pos NA si sit 2 sharp inflected2 pos pos NA sit bi 2 broad inflected1 smalUzero no dent1 any bi pos bm 2 broad monotonic1 pos 2si lbm bm The logic indicated in the table will resolve all the shapes in a static or dynamic case except for the sit case, which relies for detection on the magnitude of a positive output relative to another positive output. During the event, a small sit output can look like an sil output from a single dent that is very large, so additional classification means are required.
The si2/si 1 problem can be resolved by using a third sensor that saturates locally at a critical level of positive curvature and tends to ignore negative curvature. For this sensor, deepening 2o sit or si 1 shapes will saturate at the same depth, causing the output to stop increasing at the same time. At (or after) that point in time, the magnitude of output from a NON-saturating sensor like no. 2 above can be used to infer whether it is an sit or si 1 event. The sit event will always have a larger magnitude at sensor 2 when sensor 3 saturates, because it is like two sil outputs added.
Another means of resolving the si2/sil problem is to use a coupled lenticular sensor. Such a sensor can be made to have a very small output for an sit event, and a very large output for an si 1 event, thereby resolving the problem with great simplicity. This completes our presentation of the present example.
Figure 25 illustrates a coupled fiber sensor structure that, as in Figure 24, includes a detector of coupled light and a detector of light attenuated within the fiber connected directly to the light source. The structure of Figure 25 further includes multiple bends so that the fibers may traverse the sensing region multiple times, and so that the source and detectors may be co-located. With such a system, it is possible to use different enhanced coupling treatments on each traverse, so that the net signals are influenced by a combination of treatments. This amounts to a form of optical computer, wherein the optical signals are combined to infer shape information.
Figure 21 shows a sinuated lenticularly coupled fiber sensor designed to provide a signal that decreases as it becomes covered by liquid. It may also be used to indicate the extent of contact with a solid surface or surfaces. The throughput of the sensor of Figure 21 may be calculated according to Equation 2. If the sinuations are equal and of moderate curvature, then the sensor in air will tend to have a net output that varies little with the number of sinuations, as indicated in the curve for Ai= 0.7 in Figure 20. As each sinuation becomes covered by liquid or contacts a solid, its coupling is decreased to near zero, so the sum term of Equation 2 is decremented by a single Ki as each sinuation is contacted. This results in a linear decrease in Ec as liquid or solid contact increases in extent, falling to near-zero throughput for total contact with all the sinuations. This is to be contrasted with a single fiber version of the sensor, with source and detector on the same single fiber. The throughput of such a sensor is determined by the internal reflection conditions at each curve. From experiments with individual curves, we know that the maximum attenuation that may be achieved at a single curve that has a covering of clear material to protect it is 33%. Thus, a single fiber sensor with successively immersed curves would be expected to have a throughput that varies in steps, the net throughput having values like 1.0, .67, 0.45, 0.30, 0.20, 0.14, etc. if the throughput for no contact is normalized to 1Ø This would be a useful sensor, except that it is very difl'lcult in practice to achieve a 33% attenuation consistently. The attenuation value is highly dependent on the integrity of contact with the covering layer and the thickness and microscopic shape of the layer. This makes it difficult to form lookup tables in software to deal not only with the power law of the stepped attenuation function, but also with variability in each attenuation, which typically leads to attenuation values for single loops that vary from 10% to 33%. In contrast, the lenticularly coupled structure has a throughput that changes in equal steps down to near zero throughput, with each step dropping by 1/n where n is the number of sinuations. This is because frustration is virtually complete at each curve and the throughput is responding to the summation term of Equation 2. This evenly stepped behavior is little affected by small changes in the thickness or shape of the lenticular structure, or by contamination on its surface. Also, the small remaining throughput after total immersion can be measured to classify the medium contacting the sensor, according to its index of refraction. Typically, water with an index of 1.33 will produce approximately twice the residual throughput as hydrocarbons, with indices typically in the 1.4 to 1.5 range.
The lenticularly coupled sensor structure is also useful for forming arrays where each member of the array is a lenticularly coupled sensor with either a single curve to detect contact at a point, or multiple curves to detect progress of a contact front along the curves of the member until a near-zero throughput is achieved and the next member begins responding to contact. It may be modified, for curves too gradual to have significant coupling due to microbending, by emphasizing loss and collection by purposely forming "loss" zones at the curved portions. An array of single point contact sensors formed from lenticularly coupled sensors each with a single curve and a single light detector, will produce very large changes at each detector, typically 90%, as the member associated with the detector comes into full contact with liquid or solid. If n sinuations replace the single sinuation of a point sensor, each detector will see changes that are approximately 1/n for contact with each sinuation. Thus it is possible to form arrays with members that exhibit either binary or quasi-continuous changes in throughput, each member having near-zero throughput for total immersion, with a small residual value indicating the type of medium present according to its index of refraction.
Figures 26, 27 and 28 show two lenticularly coupled fibers from three different views. The curves in the fibers are of short radius, as the fibers are mounted in the end of tubing. The curves; or loops, 64 and 66, are adjacent and covered with a thin layer of optically transparent material, i.e., a clear epoxy. Both fibers are shown cut off short on one side of the loop, although that end may also be left uncut without consequence. The other, longer side is directed toward a light source or detector. As illustrated in Figure 26, a ray that is not within the plane of its loops is shown propagating upward in a first loop, where it exits the first loop near the apex. Although not explicitly shown in Figure 26, the ray exits at an angle directed toward the second loop or bend.
As shown in Figure 26, the result of this transfer is shown by the downwardly directed arrows in the second fiber. Figures 27 and 28 indicate the out-of plane egress of such rays near the apex of the first loop or bend of Figure 26, and their re-entry into the second loop or bend, which involves an intermediate internal reflection from the optically transparent covering.
Vertical arrows near the bottom of Figures 27 and 28 indicate the general~overall direction of light within each loop, not specific mode angles. Not all of the light exits the loop or bend, but portions traveling around and past the loop or bend are, for simplicity, not shown in the figures.
In a preferred embodiment for forming individual point sensors for a liquid sensing array, loop or bend radii approximately the diameter of the fiber are obtained by wrapping 0.25 mm diameter fibers tightly around the edge of 0.25 or 0.125 mm metal substrate.
Other materials for use as substrates include other rigid elements such as polyester or glass suitable for use in the medium to be tested. For other sensors, such as array elements with multiple sinuations, loops or bends with larger diameters may be desirable, for instance to achieve an attenuation to produce a desired result from Equation 2. For other sensors such as a side impact shape sensor, the fibers may be straight initially.
Figure 29 illustrates two elements of an array of paired loops or bends 78, built on the edge of a band of spring steel 80. In a typical array, the spring steel is 0.125 mm thick and 12.5 mm wide, and the fibers are 0.25 mm in diameter. As shown, the left-most loop or bend of each pair carnes light along the back side of the steel until it crosses over to the second loop or bend at the edge of the substrate. The light then travels along the second fiber along the front of the steel, toward a photodetector. The first loop or bend passes over the part of the second loop or bend at the back of the steel. During construction, the loops or bends are pulled tight so that the fibers touch the metal virtually everywhere along their lengths and are snug against each other and against the steel. If the steel band is narrow as shown, the natural curves of the fibers prevent orienting the long axes of the loops or bends perpendicular to the long axis of the steel, but this does not affect performance. The important factor is to achieve snug contact between fibers and to the metal. This occurs naturally, aided by the crossover of fibers and the tendency of the leads to both be placed in compression when the loops or bends are pulled tight. The loops or bends on the steel are covered in clear epoxy or a similar clear film, and in fact the entire assembly is normally covered in epoxy. Only the loop or bend apexes need remain optically clear. The rest of the assembly can be covered with opaque materials.
While they are curing, but still flowable, epoxy or similar clear liquids naturally form the correct shape for transfer of light from one loop to the other. The requisite shape is lenticular, in that it follows the curve of the loops in one dimension, and is nearly flat or an outwardly convex dome shape between the loops. This shape is ideal for the three dimensional path taken by light transferring from one loop to the other. Light exits the first loop along its length, and is reflected by a curved length of the lenticular surface, with geometry well matched to the curved length of the second loop. Thus, light that exits the first loop at multiple points is very likely to enter the second loop in a geometrically symmetrical fashion. This result is evident in the high throughput of these sensors when exposed to air.
Natural liquid forces such as capillary action cause uncured epoxy to flow in between the loops and to form a thin covering near the apexes. If optical throughput is observed during curing of the epoxy, it will be seen to improve during the initial part of the cure, when the epoxy is still capable of flowing. This is in contrast to coating two adjacent fibers cut square in the same plane. In that case, there is no transfer of light from one to the other, as the geometry deteriorates as the epoxy gets thinner on the cut faces of the fibers. For the cut fibers, there is simply not enough material on top of the cut faces for reflections to occur from one fiber to the other. The only cure would be to add a separately formed lens or reflective structure.
The clear covering material is curved in one dimension, following the curved contour of the loops. This is a desirable shape, because it creates multiple reflection paths for the light 3o emitted from the first fiber along a length of the loop. The curved shape is optimal for transferring light into the second fiber, which bears a symmetrical shape relationship to the first fiber.

Single loop pairs as shown in Figures 26, 27 and 28 may be formed by bending the fibers into tight loops and pushing them back into surrounding tubing. The end can then be dipped in epoxy. At the non looped ends of the fibers, an LED or other light source is attached to one fiber for illumination, and the other fiber is attached to a photodiode and amplifier or other similar photodetection system. The cut end of each fiber near the sensing loops may be of any length, and can be extended to provide other signaling functions or to create other loop structures along the same fiber. Normally, however, it is cut 5 to 10 mm away from the loop.
If desired, it may be covered with opaque material to prevent ingress or egress of light.
An array according to the present invention may be used in conjunction with other devices.
The optical sensor may be instrumented by attaching at least one fiber (a "first" fiber) from each pair to a light source, and the other fiber (the "second" fiber) to an individual photodetector.
An array may also be multiplexed. For example, according to the above, an array may be used in a multiplexes whereby multiple first fibers are attached to each of several light emitting diodes (LEDs), and multiple second fibers are attached to each of multiple photodetectors.
The fibers are arranged so that, for instance, four fibers from the first four looped or bent pairs of the array are illuminated by a first LED and the second looped or bent pair mates are read out by 4 photodetectors. The same 4 photodetectors are used to read out other pairs when they become illuminated by turning off the first LED and turning on another.
This system may be extended to multiplex any number of loops. A typical multiplexes is arranged to have 6 LEDs and 8 photodiodes, with 8 fibers at each LED and 6 fibers at each photodiode, for a 6 X
8 = 48 element array. Alternatively, all loop or bend pairs may be illuminated by a common source, and read out by a television camera such as a charge-coupled-device (CCD) camera or a line scanner.
An advantage of an array of discrete point sensors is the absolute accuracy with which the location of each loop is known along the substrate. When liquid first contacts a sensor pair, its location can be known with great accuracy, However, the position of liquid between point sensing pairs in an array is not known. This may be resolved by using another, continuous sensing means in conjunction with the array. The result can be a very accurate sensor combination. For example, a tank instrumented with a conventional pressure sensor has an approximate range of 1% accuracy over the range of pressures due to changes in tank level. By combining the pressure sensor with an array of 16 optical point sensors according to the present invention, spaced equally over the height of the tank, the accuracy can be improved to as good as 1/16%, using a computer to re-calibrate the pressure sensor automatically every time the liquid level passes the accurately known position of one of the optical sensors. Similarly, 48 optical sensors could be used to obtain an overall accuracy of 1/48% = 0.02%.
As illustrated in Figure 31, the array may also be used in conjunction with a standpipe inside the tank open at the bottom of the tank, and a means of varying the pressure locally within the standpipe to change the height of liquid within it. The control of local pressure requires only a small added pressure, as one need only vary the height by one inter-sensor distance. By reading the pressure over a span of one intersensor length of the array, combined with knowledge of the liquid location to the nearest intersensor interval, an instrumentation system can determine the actual liquid height before pressurization with excellent accuracy. For example, with the provision of a 48 element array and a 1% pressure sensor, an accuracy of 1/48 percent is easily achieved over the total height of the tank.
If desired, rather than use pressure to displace the liquid, one can also move the array up and down by known amounts to read the exact height of the liquid. A major advantage is that the array need not be moved by more than one intersensor length to determine the liquid height within the entire height of the tank. For instance, if there are 48 elements to the array, and the tank is 48 feet tall, there is no need to move the array more than 1 foot to determine the liquid height to great accuracy.
An array of lenticularly coupled fiber sensors, each of which has multiple sinuations may also be used to obtain highly accurate measurements of liquid height. Each member of the array can be made to have a throughput that decreases by 1/n each time liquid covers one of the n sinuations in each member so that the member has a throughput near zero when fully covered.
An array of 48 members, each with 10 sinuations, can have an absolute accuracy of 0.2 percent.
In an alternative embodiment, the sensor may be used as a humidity sensor. In use one may detect the humidity in one's breath by breathing on the loops. As such, the device may be used as a small, rapid all optical humidity sensor substitute for a chilled mirror humidity sensor.
Traditionally, one would chill the mirror to detect dew-point. By chilling the loops one would be able to detect dew-point.
In a further alternative embodiment, a lenticularly coupled sensor with pre-formed curves that 1 o couple light between the fibers, may be used as a pressure sensor.
According to the present alternative embodiment, the optical sensor includes a pair of fiber optic fibers, having a film, i.e., plastic or the like, placed against the loop or bend lens and pressure is then exerted on the film. Preferably the plastic film may be clear, colored or dark, and may even be opaque. Since the film is to some extent deformable, it will act as a fiustrator (having an index higher than air, or in the case of dark tapes, simply an absorber) whose contact area varies with pressure.
Performance is not affected by thickness of the contacting film. For example, films like 10 mil polyethylene, 4 mil mylar, 1 or 2 mil Scotch tape, black or colored vinyl tape and the like all produce similar results. This present alternative embodiment is a true index-based fivstrator, not affected by light or dark colors on the other side of the film from the loops. A linear array 20 of pressure sensors built according to this alternative embodiment, with a continuous sheath of flexible plastic between it and surrounding liquid, could be used to sense progress of the liquid along its length, according to the array members contacted by the plastic as the liquid advances, pushing the plastic against the members.
In a further alternative embodiment, a lenticularly coupled sensor without pre-formed curves may be placed between two flexible indenting plates, such as waffle-patterned rubber sheets.
Pressure applied to the sheets will cause bending of the fibers and thus coupling of light between the fibers. The throughput of the sensor will be a measure of the applied pressure or force, and can be used to classify impressed pressure pattern shapes according to the 3o curvatures imposed and the characteristic response designed into the sensor by various methods of creating enhanced coupling zones.

In a fixrther alternative embodiment shown in Figure 32, a pressure sensor array may be formed from multiple parallel fibers (fiber optic ribbon cable), by forming lenticular coupling regions 100 between adjacent fibers. If one coupling region per pair of fibers is formed, at a known position along the fibers, then the array may be used to sense magnitude and location of imposed pressure fields. Each coupling region may be formed by applying a clear flexible material so that it forms into a lenticular shape during curing, as explained previously.
Coupling at lower curvatures rnay be enhanced by creating loss and collection zones under the lenticular structure. If the array is sandwiched between flexible indenting plates, such as waffle-patterned rubber sheets, applied pressure will cause the fibers to bend and to couple light wherever a bend falls on a coupling region. Light sources and detectors may be placed at opposite ends of the ribbon, or reflectors may be applied to one end of the fiber ribbon, and all sources and receivers may be located at the other end. If reflectors are applied at one end of the cable to both receiving and transmitting fibers, then each coupling zone will couple direct and reflected light, resulting in a larger throughput.
In a further alternative embodiment, a device including the paired optic fibers and lens would be quite sensitive to chemically activated gels or the like. If desired, a sensor could be used to allow for the detection of chemicals, for use as a chemical or biological activity detector or the like.
Additionally, the device in accordance with the present invention could include a formed lens constructed from a material including dissolvable substances, such as a meltable wax, hot glue or the like. Such a sensor would be adapted to detect high temperatures or have the lenses dissolve in the presence of solvents.
In alternative embodiments, the device in accordance with the present invention could include a single fiber shape sensor for air bag deployment decisions, including whether or not to deploy, at what pressure to deploy, based on a class of shape of an object striking a portion such as the front, sides or rear of a vehicle.
In a further alternative embodiment, the device in accordance with the present invention could include a single fiber shape sensor for determining a seat occupant weight, position and shape for purposes of airbag deployment decisions, including detection of an occupied child safety seat. Decisions for any of the above embodiments could be based on suitable methods and or programs, for example, algorithms in an electronic control system of a vehicle.
In an additional alternative embodiment, the device in accordance with the present invention could include a single fiber shape sensor installed in a window, door or tailgate gasket or positioned in another suitable location, in order to detect if a hand or other body part is present. If such a body part is detected, the closure of the door, window or the like member would be interrupted.
In a fi~rther preferred alternative embodiment, the device in accordance with the present invention could include a single fiber shape sensor for detecting contact and the shape of contact between a car bumper or other vehicles, i.e. cars, trucks, construction vehicles, front end loaders, boats, boat bumpers, loading docks, marine docks and other suitable surfaces where such a sensor would desirably be placed for detecting contact and shape of contact.
In another alternative embodiment, the device in accordance with the present invention could include a single fiber shape sensor for use in an alarm system, for example as an intrusion alarm on a threshold, under a rug, or other like object, or in a window or door structure In an alternative embodiment, the device in accordance with the present invention could include a single fiber shape sensor for a safety mat, to actuate or deactuate a machine when a person steps on or offthe mat.
In an alternative embodiment, the device in accordance with the present invention could include a single fiber shape sensor for use as a pressure detector buried in pavement or on pavement, to detect and measure vehicle wheel presence, shape, speed and numbers.
Alternatively, according to the above, the single shape sensor could be used to detect, in a tire, under or over inflation.
In a further alternative embodiment, the device in accordance with the present invention could include a single fiber shape sensor for use in a bed or chair, for detection of occupant position, weight, shape and other data for purposes of position adjustment, patient monitoring or sleep research.
In an alternative embodiment, the device in accordance with the present invention could include a single fiber shape sensor for use to detect contact and the force of contact in a target, such as a gaming target, gaming tool or military target. This would allow for measurement or detection of contact and force of contact of a ball, projectile or other like device.
In an alternative embodiment, the device in accordance with the present invention could 1 o include a single fiber shape sensor for use in safety research, such as in or on the deformable elements of a crash test dummy, i.e., in or on the deformable abdomen, chest or head of a crash test dummy, to measure shape and severity of an impact.
In an alternative embodiment, the device in accordance with the present invention could include a single fiber shape sensor in an elevator to detect the presence of an obstroction between the doors or closing strips of the elevator doors.
In a further alternative embodiment, the device in accordance with the present invention could include at least one fiber in a shape sensor adapted to detect the shape of frontal impacts for 20 purposes of airbag deployment.
As those skilled in the art will realize, these preferred illustrated details can be subjected to substantial variations, without affecting the function of the illustrated embodiment. Although embodiments of the invention have been described above, it is not limited thereto and it will be apparent to those skilled in the art that numerous modifications form part of the present invention insofar as they do not depart from the spirit, nature and scope of the claimed and described invention.

Claims (46)

1. An optical sensing device comprising:
a first optical lightguide;
a second optical lightguide with a portion of its length parallel to and in close proximity to said first lightguide to form a coupling region;
said first and second lightguides being covered within the coupling region;
and wherein, light is coupled from said first lightguide to said second lightguide, when the lightguides are curved out of their plane within the coupling region.

2. A device as claimed in claim 1, wherein said cover for said coupling region is formed by a lens layer of optically transmissive material having a convex, arcuate outer surface.

3. A device as claimed in claim 1, wherein the first and second lightguides are modified for enhanced coupling, by abrasion, chemical treatment, heat forming, or notching, to lose and collect light in adjacent surface areas facing away from the plane of the lightguides in the coupling regions.

4. A device as claimed in claim 1, wherein the lens layer is formed on only one side of the plane of the lightguides, thereby enabling coupling only for curvatures of the lightguides which impose convex curvature on the lens layer.

5. A device as claimed in claim 1, means for injecting light into said first lightguide, and means for detecting the intensity of light coupled into said second lightguide.

6. A device as claimed in claim 1, means for injecting light into said first lightguide, means for detecting the intensity of light coupled into said second lightguide, and means for detecting the intensity of light carried through said first lightguide.

7. A device as claimed in claim 2, wherein the lens layer is formed on only the side of the plane of the lightguides containing the loss and collection areas, thereby enabling coupling only for curvatures of the lightguides which impose convex curvature on the lens layer.

8. A device as claimed in claim 1, said transparent material comprising a synthetic resin.

9. A device as claimed in claim 1, said transparent material comprising a heat dissolvable material.

10. A device as claimed in claim l, said transparent material comprising a chemically removable material.

11. A device as claimed in claim 1, said first and second lightguides formed into curves out of the plane of the lightguides, within the coupling region.

12. A pressure or shape measuring and classifying sensor as claimed in claim 1, said first and second lightguides mounted on a surface to be deformed by imposed pressures or shapes.

13. A pressure or shape measuring and classifying sensor as claimed in claim 12, wherein the lightguide diameter and coupling enhancement means are chosen to produce coupled light intensity that is maximal for only a single inflected shape and is attenuated when more than a single inflected shape is imposed.

14. A pressure or shape measuring and classifying sensor as claimed in claim 12, wherein the lightguide diameter and coupling enhancement means are chosen to produce coupled light intensity that is maximal for shapes with large curvatures and minimal for shapes with minimal curvature.

15. A pressure or shape measuring and classifying sensor as claimed in claim 12, wherein the lightguide diameter and coupling enhancement means are chosen to produce coupled light intensity that decreases for noninflected shapes and increases for inflected shapes.

16. A pressure or shape measuring and classifying sensor as claimed in claim 12, wherein the lightguide diameter and coupling enhancement means are chosen to produce coupled light intensity that is minimal for inflected shapes and maximal for noninflected shapes.

17. A pressure or shape classifying sensor comprising the first or second lightguide of claim 1 or claims 13 through 16 wherein the intensity of light that has passed through said lightguide is measured to classify the shape imposed on said lightguide according to the number of inflected curves, polarity of curvature, and magnitude of curvature.

18. A pressure or shape measuring and classifying sensor comprising:
a first plurality of sensors as claimed in claim 1, exposed to a distribution of curvature within an extent;
a second plurality of sensors as claimed in claim 15, exposed to a distribution of curvature within said extent;
wherein the measurements of a pressure or shape distribution by said sensors are analyzed singly and in combination to classify said distribution of curvature within said extent according to absolute value, polarity, number of inflections, number of peaks, spatial frequency content, and location within said extent, and to measure the time progress of said classifications.

19. A pressure or shape measuring and classifying sensor as claimed in claim 18, for determining classes and growth of impacted shapes in vehicles for purposes of airbag deployment.

20. A pressure or shape measuring and classifying sensor as claimed in claim 18 for determining occupant position and weight in vehicles for purposes of airbag deployment.

21. A pressure or shape sensing array comprising:
sensors with coupling regions as claimed in claim 1 distributed over an area within which pressure or shape is to be measured at locations;
wherein said sensor coupling regions are located to respond uniquely to pressure or shape at said locations;
wherein the overall pressure or shape is inferred from the individual sensor measurements.

22. A pressure or shape measuring and classifying sensor array as claimed in claim 21, wherein the sensors comprise electrical conductors instead of lightguides, said coupling regions comprise electric coupling regions wherein coupling is modulated by bending, and said bending is determined by measuring electric current or voltage resulting from said coupling.

23. A pressure or shape sensing array as claimed in claim 21, wherein said sensors are formed from adjacent fiber pairs of a fiber optic ribbon cable, wherein each coupling region occupies a known location along the axial extent of said cable.

24. A sensor as claimed in claim 1, 21 or 23 located between first and second mechanical layers, said mechanical layers containing structures capable of bending said sensors when pressure is applied.

25. A liquid or solid contact measurement sensor comprising the array as claimed in claim 21, said coupling regions preformed into curves that couple light maximally when surrounded by a medium of low index of refraction and which couple light minimally when surrounded by a medium of high index of refraction.

26. A liquid or solid contact measurement sensor comprising a sensor as claimed in claim 1, with coupling regions preformed into curves along its extent, each curve of which couples light maximally when surrounded by a medium of low index of refraction and which couples light minimally when surrounded by a medium of high index of refraction.

27. A liquid or solid contact measurement sensor comprising the array as claimed in claim 25 in which a flexible surrounding material containing air at atmospheric pressure within is deflected by pressure from a liquid or solid medium without, to touch said curved coupling regions and produce changes in the measured intensity of light indicative of contact.

28. A sensor as claimed in claim 25, 26 or 27, including a planar support member having an edge, said coupling regions spaced apart along and extending over said edge.

29. A sensor as claimed in claim 28, said coupling regions extending over said edge.

30. A liquid or solid contact measurement device as claimed in claim 1, said coupling region preformed into a curve with its apex exposed at the end of a tube covering the device.

31. A liquid contact measurement device as claimed in claim 25 or 30, the intensity of coupled light when the device is immersed in liquid indicating the index of refraction of the liquid.

32. A liquid or solid contact measurement sensor as claimed in claim 25, whereby said array is adapted to indicate the level and composition of layered liquids.

33. A liquid or solid contact measurement sensor as claimed in claim 25, comprising an array with spaced sensors, and motive means for changing the liquid or solid level with respect to the sensor array by a known displacement up to one intersensor spacing, said array measurement and said displacement being used to determine liquid or solid height or composition along a continuum.

34. A method of sensing a pressure or shape comprising the steps of:
providing a first optical lightguide;
providing a second optical lightguide with a portion of its length parallel to and in close proximity to said first lightguide within a coupling region;
covering said first and second lightguides within the coupling region, as a unit, by a lens layer of optically transmissive material having a convex, arcuate outer surface;
transmitting light from a light source through said first optical lightguide;
measuring the intensity of light coupled to said second lightguide through the lens layer, by measuring its intensity at the end of said second lightguide toward which said coupled light is directed, as a means of measuring curvatures within the coupling region.

35. A method of sensing liquid or solid contact comprising the steps of:
providing a first optical lightguide;
providing a second optical lightguide with a portion of its length parallel to and in close proximity to said first lightguide within a coupling region;
covering said first and second lightguides within said coupling region, as a unit, by a lens layer of optically transmissive material having a convex, arcuate outer surface;
forming said coupling region into at least a single curve;
transmitting light from a light source through said first optical lightguide;
measuring the intensity of light coupled to said second lightguide through the lens layer, by measuring its intensity at the end of said second lightguide toward which said coupled light is directed, as a means of measuring the contact of liquid or solid and the index of refraction of said liquid or solid.

36. An optical sensing device comprising:
an optical lightguide;
an actuation operably device associated with said optical lightguide;

said optical lightguide when deformed forming a coupling region adapted to transmit light along its length when the lightguide is curved out of its plane to said actuation device.

37. A device as claimed in claim 36, including means for injecting light into said lightguide, and means for detecting the intensity of light coupled into said lightguide.

38. A device as claimed in claim 36, wherein there is provided a cover for said coupling region formed by a lens layer of optically transmissive material having a convex, arcuate outer surface.

39. A device as claimed in claim 36, wherein said lightguide is formed into curves out of the plane of said lightguide, within the coupling region.

40. A device as claimed in claim 36, wherein said lightguide is mounted on a surface to be deformed by imposed pressures or shapes.

41. A device as claimed in claim 36, for determining classes and growth of impacted shapes in vehicles for purposes of actuating an airbag actuation device.

42. A device as claimed in claim 41, for determining occupant position and weight in vehicles for purposes of airbag deployment

43. A device as claimed in claim 36, for determining classes and growth of impacted shapes along gaskets

44. A device as claimed in claim 36, wherein said lightguide is mounted in a surface to be deformed by imposed pressures or shapes for purposes of determining classes and growth of impacted shapes.

45. A device as claimed in claim 1, wherein said lightguides are mounted in a surface to be deformed by imposed pressures or shapes for purposes of determining classes and growth of impacted shapes.

46. A pressure or shape measuring and classifying sensor as claimed in claim 1, for determining classes and growth of impacted shapes in vehicles for purposes of airbag deployment.