Proceedings of IMECE
2003 International Mechanical Engineering Congress and R and D Expo
November 16-21, 2003, Washington, D.C. USA
IMECE2003-41598
AN INVESTIGATION OF VIBRATION FEEDTHROUGH AND FEEDTHROUGH
CANCELLATION IN JOYSTICK CONTROLLED VEHICLES
R. Brent Gillespie∗
Haptics Laboratory
G 034 Lay Autolab Bldg.
Department of Mechanical Engineering
University of Michigan, Ann Arbor
Ann Arbor, Michigan, 48109
E-mail: breng@umich.edu
Szabolcs Sövényi
Haptics Laboratory
1015 Lay Autolab Bldg.
Department of Mechanical Engineering
University of Michigan, Ann Arbor
Ann Arbor, Michigan, 48109
E-mail: ssovenyi@umich.edu
ABSTRACT
The inertial and applied forces acting on operators of joystick controlled machinery in moving vehicles can produce unintentional control signals through the joystick. These forces tend
to deteriorate continuous tracking performance and further, when
the machinery in control is the vehicle itself, they may lead to unstable oscillations that jeopardize that vehicle’s safe operation. In
this paper, we propose the use of a force-reflecting joystick and a
model-based controller to cancel the effects of inertia forces. Using a simple physical model of human biomechanics, we experimentally investigate the effectiveness of a cancellation controller
in stabilizing a driving task. A second experiment involving a
human subject on board a motion base investigates the ability of
the cancellation controller to improve performance in a continuous tracking task. Results indicate that the cancellation controller
enhances stability and improves tracking.
tion feedthrough or biodynamic feedthrough [1], [2] and [3]. We
are investigating biodynamic feedthrough and its cancellation in
the context of human tracking performance. We distinguish between two types of tracking task. The first we call remote control
or open loop which involves tracking of a moving target while
onboard a moving host vehicle using a piece of machinery whose
motion is independent of the motion of the host vehicle. The
motion (shock and vibration) of the host vehicle acts on the joystick through the biodynamic system of the operator, and it can
degrade tracking performance. The second type of tracking task,
which we call driving, or closed loop, involves tracking of a moving target with the vehicle itself. In the driving task, the motion
of the machinery used for tracking is the same as the motion
which acts as disturbance to tracking. In this case the operator
may be considered a pilot or driver. A loop is closed through
the biomechanics of the operator’s body, and this may result in
unstable vehicle oscillations that jeopardize the safe operation
of the vehicle. Examples of vehicles whose piloting may suffer
from vibration feedthrough include tanks, electric wheelchairs,
frontloaders, fighter jets and helicopters [4].
INTRODUCTION
Human operators onboard moving vehicles are subjected
to inertial forces due to vehicle accelerations, and these forces,
when they are coupled through the operator’s body into the control interface, induce unintentional control signals that degrade
tracking performance. This phenomenon has been called vibra-
∗ R. Brent Gillespie (corresponding author) is with the Department of Mechanical Engineering, University of Michigan, Ann Arbor, MI 48109, USA. (e-mail:
brentg@umich.edu)
The present study focuses on eliminating vibration
feedthrough. The use of a motorized, force-reflecting joystick
has been proposed for this purpose [1]. The aim is to cancel
the unintentional force caused by inertial effects with a torque
injected by a DC motor on the joystick. This is expected to improve tracking performance in both the remote control and driving cases and to improve stability in the driving case.
1
c 2003 by ASME
Copyright °
Bode magnitude plots of the pilot-vehicle transfer functions follow trends similar to those of experimental Bode plots.
The need to predict the continuous tracking performance of
pilots of vehicles and machine operators was first identified during the second World War [12]. Since that time, ever more accurate models of human tracking performance have been sought
by the military and by industry [13]. A comprehensive summary
of such models is given by Reid [14] and by McRuer [15]. The
most frequently used types of models are the structural model
and the optimal control (also known as algorithmic) model. Alternative modelling approaches also exist, e.g. the fuzzy control model [12]. The structural model evolved from McRuer’s
crossover model, and methods for measuring or estimating its parameters have been documented in numerous articles, e.g. [16].
McRuer et al. [17] describes new experimental results in the context of this model. The human operator is modeled as a linear,
time invariant system in most of the studies. The algorithmic
model uses LQR techniques and Kalman filters, and its implementations both in simulation and hardware are described in
great detail in the literature [18]. The tracking performance of
human operators has been investigated for a variety of scenarios,
including tracking tasks carried out in more than one dimension
at a time [19]. McRuer and Schmidt [20] investigated the behavior of pilots when carrying out a secondary task in addition to
tracking.
The analysis and simulation of vibration feedthrough and
feedthrough cancellation through signal processing of a joystick
controlled aircraft is presented by Verger et al. [21]. In that work,
the inertial effects acting on the pilot are estimated by an adaptive filter and they are subtracted from the control signal. The
results of an experimental study in which a joystick controlled
motion platform was used for demonstrating the solution were
published in another paper by Verger et al. [2]. As mentioned
above, to eliminate vibrations induced by inertial effects, adaptive filtering of the control signal was implemented by Verger et
al. [21], [2]. In this work, however, the cancellation was effected
by injecting a cancelling signal to the joystick output, rather than
imposing a cancelling torque on the joystick. Thus the feel of the
joystick to the operator was not affected directly. Also, this signal processing solution cuts off the high frequency components
of the control signal above 1Hz, which somewhat deteriorates the
performance of the vehicle. An acceleration feedforward control
approach that imposed a cancelling torque on the joystick using
a force-reflecting joystick was proposed by Gillespie et al. [1]
A robust controller was implemented using force-feedback by
Sirouspour and Salcudean, [3]. An alternative approach involving increased joystick damping with decreased loop gain was
proposed by Arai et al. [22].
In this paper, the use of a force reflecting joystick and a
model based controller is further investigated as a means of solving the vibration feedthrough problem. The transfer function of
the human operator from vehicle acceleration to unintentional
We consider a motion stick, a joystick with a pivot that rotates around one axis without dead-zone and without backlash,
that produces an electrical signal as a function of angular displacement. By contrast, a force stick is a rigid joystick providing
electrical signals as a function of the force imposed on it. The
latter is often used in airplanes. At this stage of our work only a
motion stick is used for the investigations.
Force feedback in manual control interfaces has been shown
to improve human/machine performance in various tasks. The
classic example of the benefits of force-feedback is the bilateral
telemanipulator, wherein force feedback from a local master manipulator carries information about the interactions taking place
between the remote slave manipulator and its environment and
thereby enables improved task performance over telerobots without force feedback. But force-feedback has also proven beneficial in the performance of vehicle control tasks. Force-reflecting
devices improve the information content of manually controlled
vehicles within virtual environments according to Repperger and
Chandler [5], enabling improved operation of the vehicle and improved performance in tracking tasks. Yuhara et al. [6] used a
structural driver-vehicle model to design a force feedback steering wheel. The added kinesthetic information improves vehicle
handling, improves lane following both in compensatory and in
pursuit control, and reduces the mental and physical load of the
driver. A force feedback joystick was used to give information
about the motion of an unmanned air vehicle (UAV) to its pilot
operating from a remote site in the work presented by Korteling
and Borg [7]. The force-feedback system improved the performance of the pilot and reduced the mental workload associated
with maneuvering a simulated UAV. Parker et al. [8] built a force
feedback system for heavy duty hydraulic machines. The system
gives a feel to the human operator for the load acting on the tip of
the boom of an excavator. The benefits of force-feedback in these
examples, however, accrue because of the improved information
about the controlled element’s behavior available to the human
operator. We are interested in benefits to be reaped by motorizing the interface device that do not involve cognitive control or
volition on the part of the human operator. Vibration feedthrough
can occur (and, with a motorized interface can be compensated)
without any participation of cognitive processing.
Vibration feedthrough has been identified as a cause of deteriorated human/machine performance and investigated in various scenarios, including force and displacement sensing joysticks scenarios. A comprehensive overview of biodynamic effects on continuous tracking performance is available in Griffin [9]. The dynamics of both motion-type and force-type joystick interfaces and the associated human-machine system was
analyzed by Hess [10], [11]. A structural pilot-aircraft model
was constructed to analyze the roll-ratchet phenomenon. This
includes a simple biodynamic feedthrough model, a continuous
tracking model, a model for manipulator-feel system dynamics and a model for vestibular motion feedback. The resulting
2
c 2003 by ASME
Copyright °
torque imposed on the joystick will be determined based on human subject tests. The proposed controller will identify this
transfer function and will impose a torque on the joystick in the
opposite sense as a function of measured vehicle accelerations.
This approach is expected to improve tracking performance and
improve safety.
The body of this paper is organized in three parts. First,
the modelling approach of the human/machine dynamics are explained. Second, the experimental and simulated demonstration
of vibration feedthrough for the closed loop case in which the
human biomechanics were modelled using a stand-in physical
second order system is presented. Also for this case, a vibration feedthrough cancellation controller was tested using a force
reflecting joystick and the stand-in model for the biomechanics.
Finally, human subject tests of tracking performance carried out
on a single axis motion platform are discussed. The degradation
of tracking performance in moving vehicles and the effectiveness of a simple cancellation controller were demonstrated by
these tests. This is groundwork for the investigation of vibration
feedthrough in the driving task, the results of which are expected
to lead to the design of a feedthrough cancellation controller.
Xp . Finally, a gain Cp multiplies the joystick angle q to produce
a command Xc imposed on the plant P(s).
..
Xv
H ui (s)
Xr
H i(s)
+_
Figure 1.
Tui
H(s)
+
Ti
T
+
J (s)
q
Cp
Xc
P(s)
Xp
Vibration feedthrough in remote control systems
A special case of this system is obtained when Ẍp = Ẍv , that
is, when the HO is subjected to the accelerations of the plant he
or she controls with the joystick. This case is the piloting task
and its block diagram is shown in Fig. 2.
s2
..
Xp
MODELLING THE HUMAN/MACHINE COUPLED DYNAMICS
Our joystick can be modelled as a second order transfer
function J(s) from the torque applied on it to angular displacement which involves its moment of inertia and a virtual return
spring and a virtual damper. The joystick angle is multiplied by
a scalar to produce the output of the virtual plant to be controlled.
The investigation is mainly concerned about the dynamics of the
joystick and the pilot, not that of the plant or the plant controller,
so the simplest plant of unity gain was selected for the experiments, thus our tracking task is zeroth order. This also results
in a more straightforward identification of the operator’s transfer
function as a controller.
Fig.1 shows the block diagram of a general system operating
in remote control mode. The human operator (HO) is characterized by the double input, single output transfer function H(s),
with the reference signal of the tracking task Xr and vehicle acceleration Ẍv as inputs and a torque T as output. The quantity the
HO intends to control by acting through the joystick is the output
of the plant, that is, the plant position Xp . The torque T acting on
the joystick J(s) has two components, one of them we call the unintentional torque Tui which is the output of the transfer function
Hui (s) describing the unintentional effects of the vehicle acceleration Ẍv on the torque T . This is a consequence of the biodynamic
(primarily inertia) forces acting on the pilot in the moving vehicle. The other component we call the intentional torque Ti , which
is the output of the transfer function describing the action of the
intentional controller Hi (s) (involving perception, cognition and
muscle action) on an error signal or other combination of Xr and
H ui(s)
Xr
+
H i(s)
_
Tui
Ti
+
H(s)
+
T
q
J(s)
Cp
Xc
P(s)
Xp
Figure 2. Vibration feedthrough in piloting tasks
The lower feedback loop we call the tracking loop, whereas the
upper feedback loop we call the disturbance loop.
Prior to carrying out experiments with human subjects, we
designed a set of experiments involving only hardware. These
hardware experiments featured stand-in inertia, damping, and
stiffness components to capture the role of the human arm and
hand in biodynamic feedthrough. In particular, a dummy mass
was attached to the end of the joystick to capture the effects of
the effective mass of the hand and arm of the pilot. A local feedback controller on the joystick DC motor was programmed to realize a rotational spring and a rotational damper. Thus the second
order system J(s) captures some of the coupled dynamics of the
joystick and human arm/hand, including inertial, damping, and
restoring force characteristics of the human operator. The goal
of the simulations and experiments carried out on this apparatus was to demonstrate the phenomenon of feedthrough and test
a feedthrough cancellation controller in the case of the piloting
task. Two experimental platforms were used, a vibration testbed
and a ride motion simulator (RMS). Experiments with the RMS
are discussed below.
Following the experiments involving hardware alone, a pilot
study was completed with a single human subject. The tracking performance of the human operator was first characterized
3
c 2003 by ASME
Copyright °
without motion disturbance and then tracking during motion disturbance was tested for the case of remote control on-board a
moving vehicle.
VIBRATION FEEDTHROUGH SYSTEM INVOLVING
HARDWARE ALONE
At first feedthrough and feedthrough cancellation were investigated in systems comprising hardware components only.
The tests and simulation carried out on the vibration test bed
served to prepare for those carried out on the RMS. The control
system used with the RMS experiments is shown in Fig. 3. The
vehicle operates in acceleration control mode, and feedthrough
is cancelled by applying a moment equal and opposite to the inertial moment on the joystick by means of a DC motor.
Figure 4.
Test apparatus with RMS and second order operator model
the same system with a feedthrough cancellation compensator in
place with simulated and experimental data, respectively.
1.5
Compensator
-md cos(q)
Tˆ
ui
1
T
+
q
J (s)
Ca
Xc
V (s)
Xv
Joystick angle, [rad]
Ti
+
Tui
Mechanics
Figure 3.
-md cos(q)
Feedthrough cancellation of local control system
0.5
0
-0.5
-1
The RMS shown schematically in Fig. 4 is capable of producing motion in six axes (three displacements, three rotations)
using a hydraulic Stewart-Gough platform. We have, however,
restricted our attention to vibration feedthrough occurring in a
single axis: lateral displacements in the direction labelled X in
Fig. 4. The angle of the joystick is denoted q. The RMS motion
controller takes acceleration commands from the control PC, and
it sends control signals to the hydraulics of the platform. It also
transmits position, velocity, and acceleration analog signals back
to the control PC. The joystick box is equipped with its own accelerometer, and it sends acceleration signal and joystick angle
data to the control PC. The control PC uses a 1kHz sampling
frequency, it calculates the RMS acceleration reference signal
and it records the joystick angle and RMS acceleration with a
sampling frequency of 100Hz. The motion of the platform is
limited to ±0.50m whereas the rotation of the joystick is limited
to ±30◦ . The parameters of the joystick are: k = 2.0Nm/rad,
I = 0.019kg · m2 , b = 0.0167Nm/(rad/s). The moment of inertia, I includes the equivalent inertia of the DC motor rotor as
coupled through the mechanical advantage (realized using a capstan drive) between the joystick and the motor.
Figures 5 through 8 present simulated and experimental time
histories of the joystick angle. Figures 5 and 7 demonstrate vibration feedthrough (no compensator in place) with simulated and
experimental data, respectively. Figures 6 and 8 show results for
-1.5
-2
-1
0
1
2
3
4
5
6
7
8
Time, [sec]
Figure 5. Feedthrough simulation
The acceleration command of the RMS was proportional to
joystick angle. The joystick/platform oscillatory motion was initiated with a torque impulse applied to the motorized joystick of
magnitude −0.2Nm and of duration 0.29sec in both the case of
feedthrough and feedthrough cancellation at t = 0s. Both in simulation and during the tests, joystick oscillation amplitude started
to grow exponentially. In case of feedthrough cancellation the responses are determined by the natural response of joystick alone.
This is because the effects of platform motion on that of the joystick are to a large degree compensated, so the joystick moves
independently of platform accelerations, as though its base was
fixed to ground. No actuator saturation occurred while the data
shown on the graphs was recorded. When the joystick virtual
spring moment is less than the friction moment at the extremity
of the oscillation, the joystick oscillation ends with an offset.
The increasing amplitude oscillations in Figures 5 (simulation) and 7 (experiment) demonstrate feedthrough and the decreasing amplitude oscillations in Figures 6 (simulation) and 8
(experiment) indicate successful feedthrough cancellation.
4
c 2003 by ASME
Copyright °
0.2
This is easily accomplished in the case of the second order operator model, since it only relies on an estimate of the unintentional
transfer function. The HO performing a task such as tracking,
however, has two inputs and one output, as shown in Fig. 1. This
section is devoted to the experimental investigation of the behavior of the coupled human-machine system theoretically and
experimentally.
To demonstrate the degradation of tracking performance due
to vehicle motion and the concept of feedthrough cancellation using a force reflecting joystick, four tests were carried out in a single axis motion platform. The first test was designed to explore
the relationship between vehicle acceleration and unintentional
torque imposed on the joystick. The pilot held a force stick in his
hand while being shaken sideways by the platform. The platform
moved according to a filtered white noise velocity control command filtered using a first order band pass filter. The high and low
pass cutoff frequencies were 0.3Hz and 5Hz, respectively. The
unintentional torque and the platform acceleration were recorded
with a sampling rate of 100Hz. The estimate of the transfer function from acceleration to unintentional torque was then obtained
by applying the method of averaged, modified periodograms of
Welch on the two time functions. The test lasted for 5 minutes.
The result can be seen in a Bode plot in Fig. 9.
0.15
Joystick angle, [rad]
0.1
0.05
0
-0.05
-0.1
-0.15
-0.2
-2
-1
0
1
2
3
4
5
6
7
8
Time, [sec]
Figure 6.
Feedthrough cancellation simulation
0.5
0.4
0.3
Joystick angle, [rad]
0.2
0.1
0
-0.1
-0.2
-0.3
40
-0.4
30
-0.5
-2
-1
0
1
2
3
4
5
6
7
20
8
Time, [sec]
10
Magnitude, [dB]
Figure 7. Feedthrough on the RMS
0.15
0.1
-10
-20
-30
0.05
Joystick angle, [rad]
0
-40
0
-50
-0.05
-60
3
10
-0.1
Figure 9.
10
2
10
1
0
10
Frequency, [Hz]
1
10
2
10
Transfer function from platform acceleration to joystick torque
-0.15
-0.2
-2
-1
0
1
2
3
4
5
6
7
The cancellation controller is supposed to imitate this transfer function, and apply a torque on the joystick as a function of
platform acceleration, as shown in the block diagram of Fig. 10.
As a first attempt, the transfer function from acceleration
to torque was modelled by the product of the distance between
the joystick pivot and the center of the palm of the pilot (0.08m)
and an equivalent mass. The latter was 0.95kg, and it was determined experimentally by adjusting the virtual mass until the
operator felt it was optimal. The cancellation torque was equal
to the product of sensed vehicle acceleration, and a constant of
8
Time, [sec]
Figure 8. Feedthrough cancellation on the RMS
VIBRATION FEEDTHROUGH IN REMOTE CONTROL
TASKS
The feedthrough cancellation method discussed in the previous section is based on the assumption that the response of
the operator to vehicle acceleration is at least in part predictable.
5
c 2003 by ASME
Copyright °
Compensator
Tˆui
Hˆ ui(s)
25
-
..
H i (s)
+
_
Ti
+
20
..
T
q
J(s)
Ca
Xc
V (s)
Xv
15
+
Tui
Biomechanics H (s)
ui
10
Magnitude, [dB]
Xr
Figure 10. Proposed solution for feedthrough cancellation
5
0
-5
0.08 ∗ 0.95. This would correspond to a −22.4dB straight line on
Figure 9.
During the last three tests a human subject carried out a pursuit tracking task with a motion stick. The goal of these tests was
to obtain an estimate for the human/machine open loop tracking
transfer function under different test conditions. The first test involved tracking a reference signal in a steady platform. The platform moved in the second test, with the cancellation controller
turned off. The platform moved and the controller was turned
on in the last test. The reference signal of the tracking tasks was
a sum of 15 sinusoids, with amplitudes decreasing with angular
frequency. The plant output was the joystick angle multiplied by
a constant gain, hence the task was again zeroth order. To reduce
the effect of joystick dynamics on the pilot’s performance, the
joystick was programmed to zero stiffness and damping. Each
test lasted for 5 minutes. The test subject saw two cursors representing the instantaneous values of the reference signal and the
plant output on a computer screen, hence the task was a pursuit
tracking task with no preview. The data was evaluated based on
the following considerations. The reference signal has a discrete
frequency content corresponding to the 15 sinusoids. The pilot
is assumed to behave linearly, so the transfer function has to be
obtained for these and only these frequencies. The method of
modified periodograms of Welch (tfe command in Matlab) does
not allow the specification of the frequencies for which to obtain
transfer function values. Therefore, first the cross-correlation
function of the input and the output, and the autocorrelation function of the output were obtained. Then the Fourier transform of
these was computed using numerical integrals, which yielded the
cross correlation spectral density of the input and the output, and
the power spectral density of the input for the specified fifteen
frequencies. The ratio of these yields the transfer function in the
frequency domain. The results are shown for the three tests in
Figures 11, 12, and 13.
Two performance metrics were used, the crossover frequency fc and the rms average of the tracking error, eRMS . These
are shown in Table 1.
The numbers indicate performance degradation as a consequence of platform motion and performance improvement due to
the cancellation controller.
McRuer’s crossover model states that the open-loop transfer function of the HO and the plant has a −20dB/dec slope in
the crossover region according to Levine [12], with a crossover
-10
-15
-20
-3
10
Figure 11.
10
-2
10
-1
0
Frequency, [Hz]
1
10
10
Open loop transfer function of tracking in steady platform
25
20
Magnitude, [dB]
15
10
5
0
-5
-10
-15
-20
-3
10
10
-2
10
-1
0
Frequency, [Hz]
10
1
10
Figure 12. Open loop transfer function of tracking in moving platform
without compensation
Table 1. Tracking performance under various test conditions
Steady
No comp
Comp
fc [Hz]
0.55
0.35
0.35
eRMS
6.6
21.8
12.9
frequency in the range of 0.5 − 1Hz in case of unpredictable reference signals. The output is delayed with respect to the input
by 150 − 300ms in case of zeroth order systems. The crossover
region is observable on all the three plots. The dots line up when
the platform is steady, while they are somewhat scattered when
the platform moves, indicating a less consistent tracking performance. Platform motion causes a peak at 2Hz, at the frequency
where the transfer function from acceleration to joystick torque
also has a peak. In the case of an unpredictable tracking reference signal this can not be a consequence of intentional tracking
activity due to its high frequency. Instead, because of the coin6
c 2003 by ASME
Copyright °
25
smallest in case of a fixed platform case (Fig. 11), and the scatter
is intermediate for the compensated case (Fig. 13). Monotonicity (or consistency) of decreasing gain as frequency increases is
at least partially restored by the compensator.
As far as the performance metrics are considered, the
changes in eRMS and fc indicate that platform motion deteriorates
tracking performance. Compensation does not change fc significantly, but it brings about a significant reduction in eRMS which
indicates that compensation aids disturbance rejection.
Two human subjects were tested for concept demonstration.
The first subject was one of the authors, who had previous experience with tracking tasks, but was not told when the compensator
was turned on. The results presented in this paper were based
on the tests with the first subject, the results obtained with the
second subject were much the same.
20
15
Magnitude, [dB]
10
5
0
-5
-10
-15
-20
-3
10
10
-2
10
-1
Frequency, [Hz]
0
10
1
10
Figure 13. Open loop transfer function of tracking in moving platform
with compensation
CONCLUSIONS AND FUTURE WORK
Vibration feedthrough has long been known to deteriorate
the tracking performance of operators in moving vehicles and to
cause unstable oscillations. The present paper summarizes a concept demonstration solving the feedthrough problem using acceleration feedforward and a simple controller. The cancellation of
unstable, closed loop vibrations was simulated and demonstrated
on a vibration test bed and on an RMS in case of a second order
operator model. Initial human subject test results were obtained.
The harmful effect of vehicle vibration on tracking performance
was demonstrated by human subject tests. The cancellation controller improved the continuous pursuit tracking performance in
a remote control task.
In forthcoming studies the torso and the arm of the operator
will be modelled as a multi-body linkage, yielding a nonlinear
system. The torque to be applied on the joystick will be calculated from the inertial properties of the body and from the motion
of the vehicle. This will yield a cancellation controller. Experiments will be used to verify the theoretical operator model and
the effectiveness of the cancellation controller. A compensator
with a third order numerator and denominator has already been
successfully implemented in human subject tests.
Further plans include carrying out more tests with untrained
human subjects realizing feedthrough and feedthrough cancellation in piloting tasks on an RMS with motion and force sticks.
Force sensing joysticks are more prone to produce unstable oscillations according to Griffin [9].
cidence with the peak mentioned, this is suspected to be a consequence of inertial effects. A stiff stick was used to obtain the
test data shown in Fig. 9, so the peak cannot be caused by joystick resonance, hence it is suspected to be due to the motion of
segments of the operator’s arm. The drop of the transfer function at 0.006Hz and at 0.01Hz also takes place during disturbed
tracking only. The correlation between a motion disturbance that
is high-pass filtered at 0.3Hz and a performance degradation at
and below 0.01Hz is yet to be explained. There is another peak
in Figure 9 between 5Hz and 10Hz. This frequency range coincides with that of resonances caused by stretch reflexes of the
hand and the arm. It may not be possible to attribute this coincidence based on a single human subject test only, nevertheless,
this can be a starting point of further investigations. The scatter
of the points in the plot becomes greater above about 15Hz. This
is to be disregarded, since sampled time functions allow us to
identify system behavior up to only about one fifth of the sampling frequency.
Several trends are evident when comparing Figs. 11 through
13 that indicate the deterioration of tracking performance by vehicle motion, and the restoration thereof by the compensator.
One important difference between the figures is the magnitude of
the transfer functions for frequencies less than the crossover frequency. Since the error signal and the output have a small phase
shift below the crossover frequency, a greater magnitude in the
open loop transfer function indicates that the closed loop transfer
function is closer to unity, that is, indicates better tracking. On
average, the magnitude data indicate the greatest magnitude in
case of a fixed platform (Fig. 11), the lowest in case of a moving platform without compensation (Fig. 12) and an intermediate
value in case of a moving platform with the controller turned on
(Fig. 13). Also, the scatter of the points is the largest in case of a
moving platform with the controller turned off (Fig. 12), indicating the least consistent tracking performance. The scatter is the
ACKNOWLEDGMENT
The authors would like to express their gratitude to their
sponsors, the Automotive Research Center and the Detroit Arsenal of the US Army Tank Automotive and Armaments Command for financial support and providing access to and expertise
for the ride motion simulator (RMS) in Warren, Michigan. We
7
c 2003 by ASME
Copyright °
would like to thank Dr. Dragan Djurdjanovic for advice on system modelling and signal processing. Also many thanks to the
Wright Patterson Airforce Base in Dayton, Ohio for lending us a
single axis motion platform.
[14]
REFERENCES
[1] R. B. Gillespie, C. Hasser, and P. Tang, “Cancellation of
feedthrough dynamics using a force-reflecting joystick,”
Proc. ASME Dynamic Systems and Controls Division,
pp. 319–326, 1999.
[2] M. Verger, A. Grunwald, and S. Merhav, “Adaptive filtering of biodynamic stick feedthrough in manipulation tasks
on board moving platforms,” Journal of Guidance, vol. 11,
no. 2, pp. 153–158, march-April 1988.
[3] M. R. Sirouspour and S. E. Salcudean, “Robust controller
design for canceling biodynamic feedthrough,” 8th International Symposium on Experimental Robotics, ISER, July
8-11, 2002.
[4] D. T. McRuer, “Interdisciplinary interactions and dynamic
systems integration,” International Journal of Control,
vol. 59, no. 1, pp. 3–12, 1994.
[5] D. W. Repperger and A. Chandler, “A study on spatially induced ”virtual force” with an information theoretic investigation of human performance,” IEEE Transactions on Systems, Man and Cybernetics, vol. 25, no. 10, pp. 1392–1404.
[6] N. Yuhara, S. Horiuchi, T. Iijima, K. Shimizu, and
N. Asanuma, “An advanced steering system with active
kinesthetic feedback for handling qualities improvement,”
Vehicle System Dynamics, no. 27, pp. 327–355, 1997.
[7] J. E. Korteling and W. v. d. Borg, “Partial camera automation in an unmanned air vehicle,” IEEE Transactions on
Systems, Man, and Cybernetics - Part A: Systems and Humans, vol. 27, no. 2, pp. 256–262, March 1997.
[8] N. R. Parker, S. E. Salcudean, and P. D. Lawrence, “Application of force feedback to heavy duty hydraulic machines,”
pp. 375–381, 1993.
[9] R. W. McLeod and M. J. Griffin, “A review of the effects
of translational whole-body vibration on continuous manual control performance,” Journal of Sound and Vibration,
vol. 133, no. 1, pp. 55–115, 1989.
[10] R. A. Hess, “Theory of roll-ratchet phenomenon in highperformance aircraft,” Journal of Guidance, Control and
Dynamics, vol. 21, no. 4, pp. 101–108, January-February
1998.
[11] R. A. Hess, “Analyzing manipulator and feel system effects in aircraft flight control,” IEEE Transactions on Systems, Man and Cybernetics, vol. 20, no. 4, pp. 923–931,
July/August 1990.
[12] W. S. Levine, The control handbook. Chapter 80: CRC
Press, 1996.
[13] J. M. Heimerl, “Driver performance model: 1. conceptual
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
8
framework.,” Army Research Laboratory, Aberdeen Proving Ground, MD 21005-5425, ARL-TR-2581, December
2001.
L. D. Reid, “A survey of recent driver steering behavior
models suited to accident studies,” Accident Analysis and
Prevention, vol. 15, no. 1, pp. 23–40, 1983.
D. McRuer, “Human dynamics in man-machine systems,”
Automatica, vol. 16, pp. 237–253, 1980.
A. Modjtahedzadeh and R. A. Hess, “A model based driver
steering control behavior for use in assessing vehicle handling qualities,” Transactions of the ASME, Journal of
Dynamic Systems, Measurement and Control, vol. 115,
pp. 456–464, September 1993.
D. T. McRuer, R. W. Allen, D. H. Weir, and R. H. Klein,
“New results in driver steering control models,” Human
Factors, vol. 19, no. 4, pp. 381–397, 1977.
S. Baron, D. L. Kleinman, D. C. Miller, and W. H. Levison, “Application of optimal control theory to the prediction of human performance in a complex task,” Air Force
Flight Dynamics Laboratory, Air Force Systems Command,
Wright-Patterson Air Force Base, OHIO 45433. AFFDLTR-69-81, 800, pp. 1–148, April 1970 - CO455 - 114-2515.
J. B. F. Van Erp and A. B. Oving, “Control performance
with three translational degrees of freedom,” Human Factors, vol. 44, no. 1, pp. 144–156, Spring 2002.
D. McRuer and D. K. Schmidt, “Pilot-vehicle analysis
of multiaxis tasks,” Journal of Guidance, vol. 13, no. 2,
pp. 348–355, March-April 1990.
M. Verger, A. Grunwald, and S. Merhav, “Suppression
of biodynamic disturbances and pilot-induced oscillations
by adaptive filtering,” Journal of Guidance, vol. 7, no. 4,
pp. 401–409, July-August 1984.
F. Arai, J. Tateishi, and T. Fukuda, “Dynamical analysis
and suppression of human hunting in the excavator design,” Proceedings of the 2000 IEEE International Workshop on Robot and Human Interactive Communication, Osaka, Japan, September 27-29 2000.
c 2003 by ASME
Copyright °