CN108255180B - A kind of intelligence manufacture environment robot and the vehicle computational intelligence driving means of delivery and system - Google Patents
A kind of intelligence manufacture environment robot and the vehicle computational intelligence driving means of delivery and system Download PDFInfo
- Publication number
- CN108255180B CN108255180B CN201810063041.3A CN201810063041A CN108255180B CN 108255180 B CN108255180 B CN 108255180B CN 201810063041 A CN201810063041 A CN 201810063041A CN 108255180 B CN108255180 B CN 108255180B
- Authority
- CN
- China
- Prior art keywords
- mobile robot
- wolf
- electricity
- workbench
- robot
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000004519 manufacturing process Methods 0.000 title claims abstract description 14
- 230000032258 transport Effects 0.000 claims abstract description 49
- 238000013528 artificial neural network Methods 0.000 claims abstract description 47
- 238000000034 method Methods 0.000 claims abstract description 23
- 241000282461 Canis lupus Species 0.000 claims description 137
- 230000005611 electricity Effects 0.000 claims description 95
- 238000012804 iterative process Methods 0.000 claims description 12
- 241000282421 Canidae Species 0.000 claims description 9
- 230000006978 adaptation Effects 0.000 claims description 8
- 230000008569 process Effects 0.000 claims description 7
- 238000000354 decomposition reaction Methods 0.000 claims description 6
- NJPPVKZQTLUDBO-UHFFFAOYSA-N novaluron Chemical compound C1=C(Cl)C(OC(F)(F)C(OC(F)(F)F)F)=CC=C1NC(=O)NC(=O)C1=C(F)C=CC=C1F NJPPVKZQTLUDBO-UHFFFAOYSA-N 0.000 claims description 6
- 238000012549 training Methods 0.000 claims description 6
- 238000012423 maintenance Methods 0.000 claims description 5
- 230000005540 biological transmission Effects 0.000 claims description 4
- 230000005284 excitation Effects 0.000 claims description 3
- 230000008676 import Effects 0.000 claims description 3
- 230000009046 primary transport Effects 0.000 claims description 3
- 230000002068 genetic effect Effects 0.000 claims description 2
- 230000008020 evaporation Effects 0.000 claims 2
- 238000001704 evaporation Methods 0.000 claims 2
- 230000001537 neural effect Effects 0.000 claims 2
- 230000003044 adaptive effect Effects 0.000 claims 1
- 230000004888 barrier function Effects 0.000 claims 1
- 230000009471 action Effects 0.000 abstract description 2
- 238000010586 diagram Methods 0.000 description 4
- 230000008859 change Effects 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000005389 magnetism Effects 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
- G05D1/02—Control of position or course in two dimensions
- G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
- G05D1/0231—Control of position or course in two dimensions specially adapted to land vehicles using optical position detecting means
- G05D1/0246—Control of position or course in two dimensions specially adapted to land vehicles using optical position detecting means using a video camera in combination with image processing means
- G05D1/0253—Control of position or course in two dimensions specially adapted to land vehicles using optical position detecting means using a video camera in combination with image processing means extracting relative motion information from a plurality of images taken successively, e.g. visual odometry, optical flow
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
- G05D1/02—Control of position or course in two dimensions
- G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
- G05D1/0212—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory
- G05D1/0221—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory involving a learning process
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
- G05D1/02—Control of position or course in two dimensions
- G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
- G05D1/0259—Control of position or course in two dimensions specially adapted to land vehicles using magnetic or electromagnetic means
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
- G05D1/02—Control of position or course in two dimensions
- G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
- G05D1/0276—Control of position or course in two dimensions specially adapted to land vehicles using signals provided by a source external to the vehicle
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Aviation & Aerospace Engineering (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Electromagnetism (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Multimedia (AREA)
- Manipulator (AREA)
- Control Of Position, Course, Altitude, Or Attitude Of Moving Bodies (AREA)
Abstract
The invention discloses a kind of intelligence manufacture environment robot and the vehicle computational intelligence driving means of delivery and system, method includes the following steps: step 1: ground guide being arranged between different operating platform, a desktop guide rail is arranged on the table；Step 2: positioned at take the desktop machine people of object workbench from take on object workbench take object designated position grab object；Step 3: mobile robot moves along ground guide and grabs object and transports to another workbench；Step 4: after grabbing object positioned at the desktop machine people for putting object workbench, be moved to put object workbench put object designated position, complete transport of the object between workbench；Step 5: in conjunction with gray scale neural network and PID neural network, establishing power quantity predicting model, action carries out decision in next step to mobile robot.By the collaborative work of desktop machine people and mobile robot, timing, the fixed point transport of automated laboratory object are completed, cooperates the setting of mobile robot charged area, realizes the round-the-clock transport of automated laboratory.
Description
Technical field
The invention belongs to robotic conveyance fields, in particular to a kind of intelligence manufacture environment robot and vehicle computational intelligence
Drive the means of delivery and system.
Background technique
For reply new round scientific and technological revolution and industry transformation, China proposes to realize made in China 2025, in made in China
In 2025, robot has been a great concern as one of ten big major fields.In face of domestic and international keen competition environment,
" intelligence manufacture " is used as a main direction by China, but is all substantially at laboratory development phase at present.Automatic industrial is real
It tests room to have many advantages, the personnel that can save execute the iterative task time, ensure safety, reduce staffs training demand, subtract
It few mistake and increases productivity and experimental precision.
The automation of industrial laboratories is a relatively new field, and development causes the huge change of workflow
Change.In current change, in order to complete relatively easy, repeated big but high required precision task in automated laboratory, move
Mobile robot is paid special attention to.In the lab, most of the autonomous of mobile robot is that passage path planning has come
At.However the unstability that has of path planning itself and route randomness greatly affected and executes fortune using mobile robot
The reliability of defeated task.Also, mobile robot breaks down when executing transport task, and transport task majority can only interrupt.This
Outside, if during transportation can not real-time perfoming robot electric quantity prediction, easily cause robot in transit because of electricity
It uses up and can not work on, to reduce conevying efficiency.Therefore, in industrial automation laboratory, there is an urgent need to a kind of machines
Device people's computational intelligence driving means of delivery efficiently realize transport task with system.
Summary of the invention
It is an object of the invention to propose a kind of intelligence manufacture environment robot and the vehicle computational intelligence driving means of delivery
With system, timing, the fixed point transport of automated laboratory object are completed by the collaboration of desktop machine people and mobile robot,
Middle desktop machine people is responsible for the short-range pick-and-place of object, and mobile robot is transported using the long range that guide rail and terrestrial reference complete object
It is defeated, cooperate the setting of guide rail and charged area, realizes the round-the-clock transport of automated laboratory.
A kind of intelligence manufacture environment robot and vehicle computational intelligence drive the means of delivery, comprising the following steps:
Step 1: ground guide being set between the workbench where shipping point of origin and terminal, one is arranged on the table
Desktop guide rail takes object designated position, temporarily takes object location and put object designated position, takes object to refer to using remote controllers transmission
It enables；
The desktop guide rail is set among workbench, described to take object designated position, temporarily take object location and put object and refer to
Positioning is installed in the two sides of workbench, described to take object designated position, temporarily take object location and put and be all provided on object designated position
It is equipped with photosensitive sensor, and the photosensitive sensor is connected with the desktop control on workbench；
The workbench temporarily takes object location edge equipped with ranging sensing receiver；
It is described to take object instruction to refer to object is put object designated position from taking object designated position to be transported to；
Step 2: taking object to instruct positioned at taking the desktop machine people of object workbench to receive, along desktop guide rail according to photosensitive sensor
The signal of acquisition from take on object workbench take object designated position grab object after, put to take the first of object workbench temporarily put it is specified
Position；
Step 3: mobile robot reception takes object to instruct, and is moved to along ground guide and object workbench is taken temporarily to put designated position
Edge, when take the distance measuring sensor of object workbench to measure distance signal satisfaction grab object apart from when, mobile robot grab object；
It is moved to along ground guide and puts the second of the object workbench edge for temporarily putting designated position, when the ranging for putting object workbench
Distance signal that sensor measures satisfaction put object apart from when, the object of crawl is put to put object workbench second and temporarily puts specific bit
It sets；
Step 4: being received positioned at the desktop machine people for putting object workbench and object is taken to instruct, be moved to along desktop guide rail and put object work
The second of platform temporarily puts designated position, and after grabbing object, the signal according to photosensitive sensor acquisition, which is moved to, puts putting for object workbench
Transport of the object between workbench is completed in object designated position；
Step 5: after mobile robot completes primary transport, according to continuous four including mobile robot current time
The electricity at moment obtains mobile robot next moment using the mobile robot power quantity predicting model based on intelligent network
Power quantity predicting value, judge whether to continue to execute transport task according to power quantity predicting value；
The electricity at continuous four moment including the mobile robot current time is to mobile robot current time
After the electricity data at continuous 500 moment inside carries out two layers of wavelet decomposition, two groups of last four moment being taken out
High fdrequency component electricity and two groups of low frequency component electricity；
The mobile robot power quantity predicting model based on intelligent network includes the moving machine based on grey neural network
Device people electricity high frequency prediction model and mobile robot electricity low frequency prediction model based on PID neural network；It is predicted in building
Used training data is mobile robot from booting to each moment electricity data during the entire process of out of service when model
High-frequency components amount and low frequency group component after carrying out two layers of wavelet decomposition；
Wherein, the mobile robot electricity high frequency prediction model based on grey neural network is with the height at continuous 4 moment
Frequency division volume and electricity is as input data, using the high fdrequency component electricity of subsequent time as output data, to grey neural network into
Row training obtains；Mobile robot electricity low frequency prediction model based on PID neural network is with the low frequency at continuous 4 moment point
Volume and electricity is trained PID neural network using the low frequency component electricity of subsequent time as output data as input data
It obtains；
Two groups of high fdrequency component electricity at last four moment and two groups of low frequency component electricity are sequentially input based on grey mind
Mobile robot electricity high frequency prediction model through network and the mobile robot electricity low frequency based on PID neural network predict mould
Type obtains two groups of electricity high frequency predicted values and two groups of electricity low frequency predicted values, with two groups of electricity high frequency predicted values and two groups of electricity
Power quantity predicting value of the accumulated value of low frequency predicted value as mobile robot subsequent time；
If the power quantity predicting value of mobile robot subsequent time is greater than 25%, mobile robot is then returned along guide rail original road
It returns, waits transport task next time；
If the power quantity predicting value of mobile robot subsequent time is less than or equal to 25%, mobile robot is driven towards by guide rail
Mobile robot charged area, and charge information is fed back into remote controllers, meanwhile, remote controllers, which are sent instructions to, to be in
The standby machine people of same guide rail is moved to the transport task initial position in respective carter, instead of the moving machine for needing to charge
Device people carries out transport task.
Further, the weight and threshold in the mobile robot electricity high frequency prediction model based on grey neural network
Value carries out optimizing acquisition using genetic algorithm；
Step A: using rainfall layer as the weight of the mobile robot electricity high frequency prediction model based on grey neural network
And threshold value, initialize rainfall layer population, and rainfall layer parameter and population is set；
The value range of rainfall layer population scale is [20,140], and the value range of river and ocean is [2,16], ocean
Number 1, minimum dminValue range be [0.025,0.15], the value range of maximum number of iterations is [200,1000], most
The value range of big search precision is [0.015,0.15]；
Step B: setting fitness function, and determine initial optimal rainfall layer and the number of iterations t, t=1；
The corresponding weight of rainfall layer and threshold value are substituted into the mobile robot electricity high frequency prediction based on grey neural network
In model, and the electricity calculated using the mobile robot electricity high frequency prediction model based on grey neural network that rainfall layer determines
Predicted value is measured, using the inverse of the mean square deviation MSE of predicted value and actual value as the first fitness function f1 (x)；
The fitness that each rainfall layer is calculated using the first fitness function, using the corresponding rainfall layer of maximum adaptation degree as
Sea, using the corresponding rainfall layer of secondary small fitness as river, remaining rainfall layer is as the streams for flowing into river or ocean；
Step C: streams is made to import river, if it find that the solution in streams is more preferable than the solution in river, then they intercourse position
It sets；
Step D: making river flow into ocean, if the solution in river is more excellent than the solution of ocean, river exchanges position with ocean, with
Final ocean is as optimal solution；
Step E: it checks whether and meets evaporation conditions: judging whether the absolute value of the difference of the adaptive value of river and ocean is less than
Minimum dmin；
If it is less, thinking to meet evaporation conditions, remove the river, and re-start rainfall, it is random to generate newly
Rainfall layer, recalculate the fitness of each rainfall layer in rainfall layer population, otherwise return step C enters step F；
The new rainfall layer number generated at random is identical as the river quantity deleted；
Step F: judging whether to reach maximum number of iterations, if reaching, output global optimum sea is corresponding to be based on grey
The weight and threshold value of the mobile robot electricity high frequency prediction model of neural network, if not up to, enabling t=t+1, entering step
C continues next iteration.
Further, the weight and threshold in the mobile robot electricity low frequency prediction model based on PID neural network
Value carries out optimizing acquisition using wolf pack algorithm is improved；
Step 5.1: using the position of individual wolf as the mobile robot electricity low frequency prediction model based on PID neural network
In weight and threshold value, initialization wolf pack simultaneously wolf pack parameter is set；
The value range of wolf pack scale is [20,250], and the value range [1,5] of the vision radius of wolf, can remember step number is
1, the value range for probability of escaping is [0.025,0.085], and the value range of maximum search precision is [0.003,0.15], maximum
The value range [400,1000] of the number of iterations；
Step 5.2: setting fitness function, and determine initial optimal head wolf position and the number of iterations t, t=1；
It is pre- that the corresponding weight in individual wolf position and threshold value are substituted into the mobile robot electricity low frequency based on PID neural network
The power quantity predicting value that model calculates is surveyed, using the inverse of the mean square deviation MSE of predicted value and actual value as the second fitness function f2
(x)；
The fitness of every individual wolf position is calculated, using the second fitness function with the corresponding individual wolf of maximum adaptation degree
Position is as initial optimal head wolf position；
Step 5.3: the first time for finding every individual wolf updates position, to update location updating individual wolf position for the first time
It sets, and position is updated with the first time of all individual wolves, update wolf pack optimal head wolf position, j=1, j indicate individual wolf position more
New number；
Position x is updated according to the first time that formula (1) calculates every individual wolf1(t), and judge the first time being calculated
Update whether position is from the new position being not up to, if reaching, repeatedly step 5.3 updates for the first time until regaining
Position calculated to obtain fitness using first update position of individual wolf if not reaching；
xj(t)=xj-1(t)+β(r)(P(t)-xj-1(t))+rand() (1)
Wherein, xj-1(t) and xj(t) -1 update position of jth and jth time of individual wolf in the t times iterative process are indicated
Update position, x0(t) initial position before indicating individual wolf iterative operation starting in the t times iterative process；β (r) is centre
Calculation amount,T and Gen is respectively indicated when the number of iterations and maximum number of iterations, and w is big
In 2 constant, β0For the maximum excitation factor, positive number is taken, r indicates individual wolf position x0(t) fitness within sweep of the eye is best
The fitness f2 (P (t)) of the companion individual wolf position and fitness f2 (x of current individual wolf position0(t)) absolute value of the difference, r
=| f2 (P (t))-f2 (x0(t))|；Rand () is a random number in [0,1]；
Step 5.4: judging that the first time of each individual wolf updates whether position meets the following conditions, carry out second of position
It updates, more new individual wolf optimal location, j=2:
Position is updated for the first time is better than its initial position in its fitness for updating position within sweep of the eye and for the first time
Fitness；
If satisfied, then enabling individual wolf find second according to formula (1) updates position x2(t), when second updates position
Fitness be better than the fitness of current individual wolf optimal location, update position as individual wolf optimal location using second,
5.5 are entered step, third time is obtained and updates position；
If not satisfied, then second of update position is identical as first time update position, individual wolf position is constant, and goes to step
Rapid 5.6, it obtains third time and updates position；
Step 5.5: found by formula (2) meet the individual wolf of condition described in step 5.4 its within the vision the
Position is updated three times, and more new individual wolf optimal location and wolf pack head wolf optimal location, j=3 enter step 5.7；
x3(t)=x2(t)+rand().v.(Gbest-x2(t)) (2)
Wherein: x2(t)、x3(t) second update position and third time of the individual wolf in the t times iterative process are indicated more
New position；GbestFor the current optimal head wolf position of wolf pack；V is the vision radius of wolf；Rand () is that one in [- 1,1] is random
Number；
Step 5.6: updating position according to the third time that formula (3) find the individual wolf for the condition described in step 5.4 that is unsatisfactory for
Set x3(t), wolf pack head wolf optimal location is updated, j=3 enters step 5.7；
x3(t)=x2(t)+escape().s.(xc(t)-x2(t)) (3)
Wherein, xc(t) it indicates in the t times iterative process, all individual wolves carry out second after updating, the mass center position of wolf pack
It sets；S is the moving step length of wolf, s≤v；Escape () is random function；
Step 5.7: will be ranked up from low to high by updated all individual wolves three times according to fitness value；
Step 5.8: n+1 individual wolf for coming front is searched for into each individual most by non-linear simple method (NM method)
Excellent position, remaining individual repeat step 5.4- step 5.6 and search for the optimal position of each individual wolf, obtain optimal wolf pack；From optimal
Global optimum head wolf position is chosen in wolf pack；
Wherein, n is random positive integer；
Step 5.9: judging whether to reach maximum number of iterations or reach maximum search precision, if reaching, output is complete
The weight and threshold value of the corresponding mobile robot electricity low frequency prediction model based on PID neural network in office optimal head wolf position；
Otherwise, t+1 is enabled, step 5.3 is gone to, carries out next iteration.
Further, sentence according to mobile robot by the terrestrial reference with unique ID of spaced set on ground guide
Whether offset mobile robot breaks down, and detailed process is as follows:
Mobile robot will return to remote controllers by the landmark information of ground guide in real time, work as mobile robot
After the information for passing back through terrestrial reference a, if at the appointed time do not pass back through the information of next terrestrial reference a+1 in range, remotely
Controller judges that current mobile robot breaks down, and according to the position of terrestrial reference a and next terrestrial reference a+1, judgement is out of order
Position where mobile robot.
Further, when mobile robot is when ground rail breaks down, remote controllers send track switch open command,
Track switch on ground rail is opened, and sends traction instruction and is extremely located at same ground rail with the mobile robot to break down
Spare moving robot, enable spare moving robot be moved to the mobile robot position region broken down, start
The electromagnetic attraction device of spare moving robot base, adsorbs the mobile robot to break down, the moving machine that will be broken down
Device people is drawn by track switch to mobile robot region to be repaired along ground guide；
It is equipped with automatically controlled track switch between adjacent ground guide, and is set between mobile robot region to be repaired and ground guide
It is equipped with automatically controlled track switch；
The spare moving robot is located at standby machine people's waiting area, and the standby machine people waiting area is located at fortune
Defeated task starting point, and be connected to ground guide.
Further, when spare moving robot reaches the last one terrestrial reference a that the mobile robot to break down is passed through
Afterwards, the vehicle-mounted binocular camera ZED for opening spare moving robot, between the mobile robot measured and broken down away from
From enabling spare moving robot close to the mobile robot to break down according to measured distance；When execution transport task
Mobile robot breaks down, and while being towed to maintenance area, remote controllers send instructions to being in for same track
The standby machine people of idle state replaces failed machines people to execute transport task.
Further, mobile robot region to be repaired is connected to charging guide rail, waits tieing up if being located at mobile robot
If repairing standby machine people's electricity in region lower than 25%, enter the charging of mobile robot charged area along charging guide rail, it is no
Then, standby machine people's waiting area positioned at transport task starting point is entered along charging guide rail.
A kind of intelligence manufacture environment robot and vehicle computational intelligence drive delivery system, including desktop machine people, desktop
Controller, mobile robot, guide rail and remote controllers；
The guide rail includes desktop guide rail and ground guide, and automatically controlled track switch, and ground are equipped between adjacent ground guide
The terrestrial reference with unique ID is arranged at intervals on guide rail；
Automatically controlled track switch on the desktop control, mobile robot and guide rail is led to the remote controllers
Letter；
The desktop machine people is set on workbench, is controlled by desktop control, and the desktop guide rail on workbench moves
It is dynamic, the fixed object location that takes, puts is provided on workbench, and take, put object location equipped with photosensitive sensor, the light sensor
Device is connected with desktop control；
Workbench is equipped with the fixed side for taking, putting object location and is provided with ranging sensing receiver；
The mobile robot is equipped with distance measuring sensor, is set to ground, adopts the ground with the aforedescribed process on ground
Face guide rail is mobile.
It further, further include mobile robot charged area and standby machine people's waiting area；
The mobile robot charged area is set to the charging guide rail between transport task initial position and end position
On, the charging guide rail is connected to standby machine people's waiting area guide rail, and the standby machine people waiting area guide rail and ground
The connection of face guide rail.
It further, further include mobile robot region to be repaired, the mobile robot region to be repaired is led with ground
It is provided with automatically controlled track switch between rail, and is provided with electromagnetic attraction device and vehicle-mounted binocular camera ZED, institute on mobile robot pedestal
Vehicle-mounted binocular ZED is stated to be arranged in above the pedestal of robot.
Beneficial effect
The present invention provides a kind of intelligence manufacture environment robot and the vehicle computational intelligence driving means of delivery and system, lead to
The timing of automated laboratory object, fixed point transport are completed in the collaboration for crossing desktop machine people and mobile robot, wherein tabletop machine
Device people is responsible for the short-range pick-and-place of object, and mobile robot completes the long-distance transportation of object, cooperation using guide rail and terrestrial reference
The setting of charged area realizes that the round-the-clock transport of automated laboratory has the advantage that in terms of existing technologies
(1) mobile robot is transported using the remote orientation that guide rail realizes object, avoids complicated path planning
The unstability of algorithm and its algorithm improves the Stability and veracity of transport；
(2) electricity based on robot has unstable and jump characteristic, establishes based on grey neural network and PID
The mobile robot power quantity predicting model of neural network carries out decision to the action next time of mobile robot, to be effectively reduced
The failure rate of transport task；
(3) it carries out optimizing using parameter of the water round-robin algorithm to grey neural network and uses to improve wolf pack algorithm pair
The parameter of PID neural network carries out optimizing, the problem of so as to avoid local convergence, and improves power quantity predicting precision；
(4) multiclass humanoid robot cooperates, and desktop machine people is responsible for the short-range pick-and-place of object, and mobile robot is responsible for
The long-distance transportation of object substantially increases the efficiency of transport.
(5) utilize the combination of ground guide and terrestrial reference, position to mobile robot and transport task carry out it is simple and
It effectively monitors in real time, provides guarantee for smoothly completing for transport task.
Detailed description of the invention
Fig. 1 is the flow diagram of the method for the invention；
Fig. 2 is the structural schematic diagram of mobile robot of the present invention；
Fig. 3 is transit domain middle orbit distribution schematic diagram in the present invention；
Fig. 4 is power quantity predicting model schematic of the present invention；
Fig. 5 is the intelligent means of delivery system diagram of the present invention.
Specific embodiment
The present invention is described further below in conjunction with drawings and examples.
As shown in Figure 1, a kind of intelligence manufacture environment robot and vehicle computational intelligence drive the means of delivery, including following step
It is rapid:
Step 1: ground guide being set between the workbench where shipping point of origin and terminal, one is arranged on the table
Desktop guide rail takes object designated position, temporarily takes object location and put object designated position, takes object to refer to using remote controllers transmission
It enables；
It takes object to instruct to mobile robot and desktop control transmission using remote controllers, mobile robot is enabled to be moved to
Object workbench is taken, enables desktop machine people be moved to and object designated position is taken to start to take object task；
The desktop guide rail is set among workbench, described to take object designated position, temporarily take object location and put object and refer to
Positioning is installed in the two sides of workbench, described to take object designated position, temporarily take object location and put and be all provided on object designated position
It is equipped with photosensitive sensor, and the photosensitive sensor is connected with the desktop control on workbench；
The workbench temporarily takes object location edge equipped with ranging sensing receiver；
It is described to take object instruction to refer to object is put object designated position from taking object designated position to be transported to；
Step 2: taking object to instruct positioned at taking the desktop machine people of object workbench to receive, along desktop guide rail according to photosensitive sensor
The signal of acquisition from take on object workbench take object designated position grab object after, put to take the first of object workbench temporarily put it is specified
Position；
Step 3: mobile robot reception takes object to instruct, and is moved to along ground guide and object workbench is taken temporarily to put designated position
Edge, when take the distance measuring sensor of object workbench to measure distance signal satisfaction grab object apart from when, mobile robot grab object；
It is moved to along ground guide and puts the second of the object workbench edge for temporarily putting designated position, when the ranging for putting object workbench
Distance signal that sensor measures satisfaction put object apart from when, the object of crawl is put to put object workbench second and temporarily puts specific bit
It sets；
Step 4: being received positioned at the desktop machine people for putting object workbench and object is taken to instruct, be moved to along desktop guide rail and put object work
The second of platform temporarily puts designated position, and after grabbing object, the signal according to photosensitive sensor acquisition, which is moved to, puts putting for object workbench
Transport of the object between workbench is completed in object designated position；
Step 5: after mobile robot completes primary transport, according to continuous four including mobile robot current time
The electricity at moment, using the mobile robot power quantity predicting model based on intelligent network, as shown in figure 4, obtaining mobile robot
The power quantity predicting value at next moment judges whether to continue to execute transport task according to power quantity predicting value；
The electricity at continuous four moment including the mobile robot current time is to mobile robot current time
After the electricity data at continuous 500 moment inside carries out two layers of wavelet decomposition, two groups of last four moment being taken out
High fdrequency component electricity and two groups of low frequency component electricity；
The mobile robot power quantity predicting model based on intelligent network includes the moving machine based on grey neural network
Device people electricity high frequency prediction model and mobile robot electricity low frequency prediction model based on PID neural network；It is predicted in building
Used training data is mobile robot from booting to each moment electricity data during the entire process of out of service when model
High-frequency components amount and low frequency group component after carrying out two layers of wavelet decomposition；
Wherein, the mobile robot electricity high frequency prediction model based on grey neural network is with the height at continuous 4 moment
Frequency division volume and electricity is as input data, using the high fdrequency component electricity of subsequent time as output data, to grey neural network into
Row training obtains；Mobile robot electricity low frequency prediction model based on PID neural network is with the low frequency at continuous 4 moment point
Volume and electricity is trained PID neural network using the low frequency component electricity of subsequent time as output data as input data
It obtains；
Two groups of high fdrequency component electricity at last four moment and two groups of low frequency component electricity are sequentially input based on grey mind
Mobile robot electricity high frequency prediction model through network and the mobile robot electricity low frequency based on PID neural network predict mould
Type obtains two groups of electricity high frequency predicted values and two groups of electricity low frequency predicted values, with two groups of electricity high frequency predicted values and two groups of electricity
Power quantity predicting value of the accumulated value of low frequency predicted value as mobile robot subsequent time；
If the power quantity predicting value of mobile robot subsequent time is greater than 25%, mobile robot is then returned along guide rail original road
It returns, waits transport task next time；
If the power quantity predicting value of mobile robot subsequent time is less than or equal to 25%, mobile robot is driven towards by guide rail
Mobile robot charged area, and charge information is fed back into remote controllers, meanwhile, remote controllers, which are sent instructions to, to be in
The standby machine people of same guide rail is moved to the transport task initial position in respective carter, instead of the moving machine for needing to charge
Device people carries out transport task.
Weight and threshold value in the mobile robot electricity high frequency prediction model based on grey neural network is using something lost
Propagation algorithm carries out optimizing acquisition；
Step A: using rainfall layer as the weight of the mobile robot electricity high frequency prediction model based on grey neural network
And threshold value, initialize rainfall layer population, and rainfall layer parameter and population is set；
The value range of rainfall layer population scale is [20,140], and the value range of river and ocean is [2,16], ocean
Number 1, minimum dminValue range be [0.025,0.15], the value range of maximum number of iterations is [200,1000], most
The value range of big search precision is [0.015,0.15]；
Step B: setting fitness function, and determine initial optimal rainfall layer and the number of iterations t, t=1；
The corresponding weight of rainfall layer and threshold value are substituted into the mobile robot electricity high frequency prediction based on grey neural network
In model, and the electricity calculated using the mobile robot electricity high frequency prediction model based on grey neural network that rainfall layer determines
Predicted value is measured, using the inverse of the mean square deviation MSE of predicted value and actual value as the first fitness function f1 (x)；
The fitness that each rainfall layer is calculated using the first fitness function, using the corresponding rainfall layer of maximum adaptation degree as
Sea, using the corresponding rainfall layer of secondary small fitness as river, remaining rainfall layer is as the streams for flowing into river or ocean；
Step C: streams is made to import river, if it find that the solution in streams is more preferable than the solution in river, then they intercourse position
It sets；
Step D: making river flow into ocean, if the solution in river is more excellent than the solution of ocean, river exchanges position with ocean, with
Final ocean is as optimal solution；
Step E: it checks whether and meets evaporation conditions: judging whether the absolute value of the difference of the adaptive value of river and ocean is less than
Minimum dmin；
If it is less, thinking to meet evaporation conditions, remove the river, and re-start rainfall, it is random to generate newly
Rainfall layer, recalculate the fitness of each rainfall layer in rainfall layer population, otherwise return step C enters step F；
The new rainfall layer number generated at random is identical as the river quantity deleted；
Step F: judging whether to reach maximum number of iterations, if reaching, output global optimum sea is corresponding to be based on grey
The weight and threshold value of the mobile robot electricity high frequency prediction model of neural network, if not up to, enabling t=t+1, entering step
C continues next iteration.
Weight and threshold value in the mobile robot electricity low frequency prediction model based on PID neural network is using improvement
Wolf pack algorithm carries out optimizing acquisition；
Step 5.1: using the position of individual wolf as the mobile robot electricity low frequency prediction model based on PID neural network
In weight and threshold value, initialization wolf pack simultaneously wolf pack parameter is set；
The value range of wolf pack scale is [20,250], and the value range [1,5] of the vision radius of wolf, can remember step number is
1, the value range for probability of escaping is [0.025,0.085], and the value range of maximum search precision is [0.003,0.15], maximum
The value range [400,1000] of the number of iterations；
Step 5.2: setting fitness function, and determine initial optimal head wolf position and the number of iterations t, t=1；
It is pre- that the corresponding weight in individual wolf position and threshold value are substituted into the mobile robot electricity low frequency based on PID neural network
The power quantity predicting value that model calculates is surveyed, using the inverse of the mean square deviation MSE of predicted value and actual value as the second fitness function f2
(x)；
The fitness of every individual wolf position is calculated, using the second fitness function with the corresponding individual wolf of maximum adaptation degree
Position is as initial optimal head wolf position；
Step 5.3: the first time for finding every individual wolf updates position, to update location updating individual wolf position for the first time
It sets, and position is updated with the first time of all individual wolves, update wolf pack optimal head wolf position, j=1, j indicate individual wolf position more
New number；
Position x is updated according to the first time that formula (1) calculates every individual wolf1(t), and judge the first time being calculated
Update whether position is from the new position being not up to, if reaching, repeatedly step 5.3 updates for the first time until regaining
Position calculated to obtain fitness using first update position of individual wolf if not reaching；
xj(t)=xj-1(t)+β(r)(P(t)-xj-1(t))+rand() (1)
Wherein, xj-1(t) and xj(t) -1 update position of jth and jth time of individual wolf in the t times iterative process are indicated
Update position, x0(t) initial position before indicating individual wolf iterative operation starting in the t times iterative process；β (r) is centre
Calculation amount,T and Gen is respectively indicated when the number of iterations and maximum number of iterations, and w is big
In 2 constant, β0For the maximum excitation factor, positive number is taken, r indicates individual wolf position x0(t) fitness within sweep of the eye is best
The fitness f2 (P (t)) of the companion individual wolf position and fitness f2 (x of current individual wolf position0(t)) absolute value of the difference, r
=| f2 (P (t))-f2 (x0(t))|；Rand () is a random number in [0,1]；
Step 5.4: judging that the first time of each individual wolf updates whether position meets the following conditions, carry out second of position
It updates, more new individual wolf optimal location, j=2:
Position is updated for the first time is better than its initial position in its fitness for updating position within sweep of the eye and for the first time
Fitness；
If satisfied, then enabling individual wolf find second according to formula (1) updates position x2(t), when second updates position
Fitness be better than the fitness of current individual wolf optimal location, update position as individual wolf optimal location using second,
5.5 are entered step, third time is obtained and updates position；
If not satisfied, then second of update position is identical as first time update position, individual wolf position is constant, and goes to step
Rapid 5.6, it obtains third time and updates position；
Step 5.5: found by formula (2) meet the individual wolf of condition described in step 5.4 its within the vision the
Position is updated three times, and more new individual wolf optimal location and wolf pack head wolf optimal location, j=3 enter step 5.7；
x3(t)=x2(t)+rand().v.(Gbest-x2(t)) (2)
Wherein: x2(t)、x3(t) second update position and third time of the individual wolf in the t times iterative process are indicated more
New position；GbestFor the current optimal head wolf position of wolf pack；V is the vision radius of wolf；Rand () is that one in [- 1,1] is random
Number；
Step 5.6: updating position according to the third time that formula (3) find the individual wolf for the condition described in step 5.4 that is unsatisfactory for
Set x3(t), wolf pack head wolf optimal location is updated, j=3 enters step 5.7；
x3(t)=x2(t)+escape().s.(xc(t)-x2(t)) (3)
Wherein, xc(t) it indicates in the t times iterative process, all individual wolves carry out second after updating, the mass center position of wolf pack
It sets；S is the moving step length of wolf, s≤v；Escape () is random function；
Step 5.7: will be ranked up from low to high by updated all individual wolves three times according to fitness value；
Step 5.8: n+1 individual wolf for coming front is searched for into each individual most by non-linear simple method (NM method)
Excellent position, remaining individual repeat step 5.4- step 5.6 and search for the optimal position of each individual wolf, obtain optimal wolf pack；From optimal
Global optimum head wolf position is chosen in wolf pack；
Wherein, n is random positive integer；
Step 5.9: judging whether to reach maximum number of iterations or reach maximum search precision, if reaching, output is complete
The weight and threshold value of the corresponding mobile robot electricity low frequency prediction model based on PID neural network in office optimal head wolf position；
Otherwise, t+1 is enabled, step 5.3 is gone to, carries out next iteration.
During taking object, according to mobile robot by the ground with unique ID of spaced set on ground guide
Mark, judges whether mobile robot breaks down, detailed process is as follows:
Mobile robot will return to remote controllers by the landmark information of ground guide in real time, work as mobile robot
After the information for passing back through terrestrial reference a, if at the appointed time do not pass back through the information of next terrestrial reference a+1 in range, remotely
Controller judges that current mobile robot breaks down, and according to the position of terrestrial reference a and next terrestrial reference a+1, judgement is out of order
Position where mobile robot.
When mobile robot is when ground rail breaks down, remote controllers send track switch open command, by ground rail
Track switch on road is opened, and sends traction instruction to the spare shifting for being located at same ground rail with the mobile robot to break down
Mobile robot enables spare moving robot be moved to the mobile robot position region broken down, and starts spare moving
The electromagnetic attraction device of robot base, adsorbs the mobile robot to break down, by the mobile robot to break down along ground
Face guide rail is drawn by track switch to mobile robot region to be repaired；
It is equipped with automatically controlled track switch between adjacent ground guide, and is set between mobile robot region to be repaired and ground guide
It is equipped with automatically controlled track switch；
The spare moving robot is located at standby machine people's waiting area, and the standby machine people waiting area is located at fortune
Defeated task starting point, and be connected to ground guide.
When spare moving robot reach the mobile robot that breaks down by the last one terrestrial reference a after, open standby
With the vehicle-mounted binocular camera ZED of mobile robot, the distance between the mobile robot measured and broken down, foundation is surveyed
The distance obtained enables spare moving robot close to the mobile robot to break down.
When the mobile robot for executing transport task breaks down, while being towed to maintenance area, remote controllers
The standby machine people being in idle condition for sending instructions to same track replaces failed machines people to execute transport task.
Robot on different tracks is numbered, the mobile robot on track 1 is named as 1A, is in spare moving
The spare moving robot of robot waiting area is named as 1B, 1C, 1D, 1E, 1F；Mobile robot on track 2 is named as
2A, standby machine human life entitled 2B, 2C, 2D, 2E, 2F in spare moving robot waiting area, track 3 and so on.
If remote server discovery mobile robot 1A breaks down when executing transport task, sends instructions to and be in
Traction failed machines people is gone before spare moving robot 1B on same track.
After spare moving robot 1B successfully draws failure mobile robot 1A to maintenance area, send instructions to long-range
Server.After remote server receives instruction, then same track spare moving robot 1C is ordered to replace mobile robot 1A, weight
It is new to start transport task.
Spare moving robot 1B returns to spare moving robot waiting area and awaits orders according to self electric quantity.
The setting charging guide rail between transport task Origin And Destination, and the guide rail that charges is connected to ground guide, is being charged
Mobile robot charged area is set on guide rail；
After mobile robot completes a transport task, mobile robot is made according to self electric quantity with making policy decision:
If the power quantity predicting value of mobile robot subsequent time is greater than 25%, along ground guide backtracking, under waiting
Transport task；
If the power quantity predicting value of mobile robot subsequent time is less than or equal to 25%, mobile robot is moved along ground guide
It moves to charging guide rail, reaches the progress of mobile robot charged area, and charge information is fed back into remote controllers.
The mobile robot region to be repaired is connected to charging guide rail, standby in mobile robot region to be repaired
If being lower than 25% with robot electric quantity, enter the charging of mobile robot charged area along charging guide rail, otherwise, along charging guide rail
Into the standby machine people's waiting area for being located at transport task starting point.
A kind of intelligence manufacture environment robot and vehicle computational intelligence drive delivery system, as shown in figure 5, including tabletop machine
Device people, desktop control, mobile robot, guide rail and remote controllers；
As shown in figure 3, the guide rail includes desktop guide rail and ground guide, automatically controlled road is equipped between adjacent ground guide
Trouble, and the terrestrial reference with unique ID is arranged at intervals on ground guide；
Automatically controlled track switch on the desktop control, mobile robot and guide rail is led to the remote controllers
Letter；
The desktop machine people is set on workbench, is controlled by desktop control, and the desktop guide rail on workbench moves
It is dynamic, the fixed object location that takes, puts is provided on workbench, and take, put object location equipped with photosensitive sensor, the light sensor
Device is connected with desktop control；
Workbench is equipped with the fixed side for taking, putting object location and is provided with ranging sensing receiver；
The mobile robot is equipped with distance measuring sensor, ground is set to, using a kind of above-mentioned intelligence manufacture environment machine
The ground guide of device people and the vehicle computational intelligence driving means of delivery on ground moves.
It is additionally provided with mobile robot charged area, standby machine people's waiting area in transit domain, mobile robot waits for
Maintenance area；
The mobile robot charged area is set to the charging guide rail between transport task initial position and end position
On, the charging guide rail is connected to standby machine people's waiting area guide rail, and the standby machine people waiting area guide rail and ground
The connection of face guide rail.
Automatically controlled track switch, and mobile robot pedestal are provided between mobile robot region to be repaired and ground guide
On be provided with earth magnetism suction device and vehicle-mounted binocular camera ZED, as shown in Fig. 2, robot is arranged in the vehicle-mounted binocular ZED
Above pedestal.
Invention is explained in detail in conjunction with specific embodiments above, these not constitute the limitation to invention.
Without departing from the principles of the present invention, those skilled in the art can also make many modification and improvement, these are also answered
It belongs to the scope of protection of the present invention.
Claims (10)
1. a kind of intelligence manufacture environment robot and vehicle computational intelligence drive the means of delivery, which is characterized in that including following step
It is rapid:
Step 1: ground guide being set between the workbench where shipping point of origin and terminal, a desktop is set on the table
Guide rail takes object designated position, temporarily takes object location and put object designated position, takes object to instruct using remote controllers transmission；
The desktop guide rail is set among workbench, described to take object designated position, temporarily take object location and put object specific bit
It installs in the two sides of workbench, it is described to take object designated position, temporarily take object location and put and be provided on object designated position
Photosensitive sensor, and the photosensitive sensor is connected with the desktop control on workbench；
The workbench temporarily takes object location edge equipped with ranging sensing receiver；
It is described to take object instruction to refer to object is put object designated position from taking object designated position to be transported to；
Step 2: taking object to instruct positioned at taking the desktop machine people of object workbench to receive, acquired along desktop guide rail according to photosensitive sensor
Signal from take on object workbench take object designated position to grab object after, put to taking the first of object workbench temporarily to put specific bit
It sets；
Step 3: mobile robot reception takes object to instruct, and the side for taking object workbench temporarily to put designated position is moved to along ground guide
Edge, when take the distance measuring sensor of object workbench to measure distance signal satisfaction grab object apart from when, mobile robot grab object；
It is moved to along ground guide and puts the second of the object workbench edge for temporarily putting designated position, when the ranging for putting object workbench senses
Distance signal that device measures satisfaction put object apart from when, the object of crawl is put to put object workbench second and temporarily puts designated position；
Step 4: being received positioned at the desktop machine people for putting object workbench and object is taken to instruct, be moved to along desktop guide rail and put object workbench
Second temporarily puts designated position, and after grabbing object, the signal according to photosensitive sensor acquisition, which is moved to, to be put the object of putting of object workbench and refer to
Positioning is set, and transport of the object between workbench is completed；
Step 5: after mobile robot completes primary transport, according to continuous four moment including mobile robot current time
Electricity the electricity at mobile robot next moment is obtained using the mobile robot power quantity predicting model based on intelligent network
Predicted value is measured, judges whether to continue to execute transport task according to power quantity predicting value；
The electricity at continuous four moment including the mobile robot current time is to including mobile robot current time
Continuous 500 moment electricity data carry out two layers of wavelet decomposition after, two groups of high frequencies at last four moment being taken out
Component electricity and two groups of low frequency component electricity；
The mobile robot power quantity predicting model based on intelligent network includes the mobile robot based on grey neural network
Electricity high frequency prediction model and mobile robot electricity low frequency prediction model based on PID neural network；In building prediction model
When used training data to be mobile robot carry out from booting to each moment electricity data during the entire process of out of service
High-frequency components amount and low frequency group component after two layers of wavelet decomposition；
Wherein, the mobile robot electricity high frequency prediction model based on grey neural network is with the high frequency division at continuous 4 moment
Volume and electricity instructs grey neural network using the high fdrequency component electricity of subsequent time as output data as input data
Practice and obtains；Mobile robot electricity low frequency prediction model based on PID neural network is the low frequency component electricity with continuous 4 moment
Amount is trained PID neural network and is obtained as input data using the low frequency component electricity of subsequent time as output data
?；
Two groups of high fdrequency component electricity at last four moment and two groups of low frequency component electricity are sequentially input based on gray neural net
The mobile robot electricity high frequency prediction model of network and mobile robot electricity low frequency prediction model based on PID neural network,
Two groups of electricity high frequency predicted values and two groups of electricity low frequency predicted values are obtained, with two groups of electricity high frequency predicted values and two groups of electricity low frequencies
Power quantity predicting value of the accumulated value of predicted value as mobile robot subsequent time；
If the power quantity predicting value of mobile robot subsequent time is greater than 25%, mobile robot then along guide rail backtracking, etc.
To transport task next time；
If the power quantity predicting value of mobile robot subsequent time is less than or equal to 25%, mobile robot drives towards movement by guide rail
Robot charged area, and charge information is fed back into remote controllers, meanwhile, remote controllers are sent instructions in same
The standby machine people of guide rail is moved to the transport task initial position in respective carter, instead of the mobile robot for needing to charge
Carry out transport task.
2. the method according to claim 1, wherein the mobile robot electricity based on grey neural network
Weight and threshold value in high frequency prediction model carry out optimizing acquisition using genetic algorithm；
Step A: using rainfall layer as the weight and threshold of the mobile robot electricity high frequency prediction model based on grey neural network
Value initializes rainfall layer population, and rainfall layer parameter and population is arranged；
The value range of rainfall layer population scale is [20,140], and the value range of river and ocean is [2,16], ocean number
1, minimum dminValue range be [0.025,0.15], the value range of maximum number of iterations is [200,1000], is most wantonly searched for
The value range of Suo Jingdu is [0.015,0.15]；
Step B: setting fitness function, and determine initial optimal rainfall layer and the number of iterations t, t=1；
The corresponding weight of rainfall layer and threshold value are substituted into the mobile robot electricity high frequency prediction model based on grey neural network
In, and it is pre- using the electricity that the mobile robot electricity high frequency prediction model based on grey neural network that rainfall layer determines calculates
Measured value, using the inverse of the mean square deviation MSE of predicted value and actual value as the first fitness function f1 (x)；
The fitness that each rainfall layer is calculated using the first fitness function, using the corresponding rainfall layer of maximum adaptation degree as greatly
Sea, using the corresponding rainfall layer of secondary small fitness as river, remaining rainfall layer is as the streams for flowing into river or ocean；
Step C: streams is made to import river, if it find that the solution in streams is more preferable than the solution in river, then they intercourse position；
Step D: river is made to flow into ocean, if the solution in river is more excellent than the solution of ocean, river exchanges position with ocean, with final
Ocean is as optimal solution；
Step E: it checks whether and meets evaporation conditions: it is minimum to judge whether the absolute value of the difference of the adaptive value of river and ocean is less than
Value dmin；
If it is less, thinking to meet evaporation conditions, remove the river, and re-start rainfall, generates new drop at random
Rain layer, recalculates the fitness of each rainfall layer in rainfall layer population, otherwise return step C enters step F；
The new rainfall layer number generated at random is identical as the river quantity deleted；
Step F: judging whether to reach maximum number of iterations, if reaching, output global optimum sea is corresponding to be based on gray neural
The weight and threshold value of the mobile robot electricity high frequency prediction model of network, if not up to, enabling t=t+1, entering step C, after
Continuous next iteration.
3. the method according to claim 1, wherein the mobile robot electricity based on PID neural network
Weight and threshold value in low frequency prediction model carry out optimizing acquisition using wolf pack algorithm is improved；
Step 5.1: in using the position of individual wolf as the mobile robot electricity low frequency prediction model based on PID neural network
Weight and threshold value initialize wolf pack and wolf pack parameter are arranged；
The value range of wolf pack scale is [20,250], and the value range [1,5] of the vision radius of wolf, can remember step number is 1, is escaped
The value range for running probability is [0.025,0.085], and the value range of maximum search precision is [0.003,0.15], greatest iteration
The value range [400,1000] of number；
Step 5.2: setting fitness function, and determine initial optimal head wolf position and the number of iterations t, t=1；
The corresponding weight in individual wolf position and threshold value are substituted into the mobile robot electricity low frequency based on PID neural network and predict mould
The power quantity predicting value that type calculates, using the inverse of the mean square deviation MSE of predicted value and actual value as the second fitness function f2 (x)；
The fitness of every individual wolf position is calculated, using the second fitness function with the corresponding individual wolf position of maximum adaptation degree
As initial optimal head wolf position；
Step 5.3: the first time for finding every individual wolf updates position, to update location updating individual wolf position for the first time, and
Position is updated with the first time of all individual wolves, updates wolf pack optimal head wolf position, j=1, j indicate individual wolf location updating
Number；
Position x is updated according to the first time that formula (1) calculates every individual wolf1(t), and judge that the first time being calculated updates
Whether position is from the new position being not up to, if reaching, repeatedly step 5.3 updates position until regaining for the first time
It sets, if not reaching, calculates to obtain fitness using first update position of individual wolf；
xj(t)=xj-1(t)+β(r)(P(t)-xj-1(t))+rand() (1)
Wherein, xj-1(t) and xj(t) indicate that individual wolf -1 update position of jth and jth in the t times iterative process time updates
Position, x0(t) initial position before indicating individual wolf iterative operation starting in the t times iterative process；β (r) is intermediate computations
Amount,T and Gen is respectively indicated when the number of iterations and maximum number of iterations, and w is greater than 2
Constant, β0For the maximum excitation factor, positive number is taken, r indicates individual wolf position x0(t) best same of fitness within sweep of the eye
Fitness f2 (the x of fitness f2 (P (t)) and current individual wolf position with individual wolf position0(t)) absolute value of the difference, r=
|f2(P(t))-f2(x0(t))|；Rand () is a random number in [0,1]；
Step 5.4: judging that the first time of each individual wolf updates whether position meets the following conditions, carry out second of position more
Newly, more new individual wolf optimal location, j=2:
The adaptation that position is better than its initial position in its fitness for updating position within sweep of the eye and for the first time is updated for the first time
Degree；
If satisfied, then enabling individual wolf find second according to formula (1) updates position x2(t), the adaptation of position is updated when second
Degree is better than the fitness of current individual wolf optimal location, updates position as individual wolf optimal location, into step using second
Rapid 5.5, it obtains third time and updates position；
If not satisfied, then second of update position is identical as first time update position, individual wolf position is constant, and goes to step
5.6, it obtains third time and updates position；
Step 5.5: finding the individual wolf for the condition described in step 5.4 that meets in its third time within the vision by formula (2)
Position is updated, more new individual wolf optimal location and wolf pack head wolf optimal location, j=3 enter step 5.7；
x3(t)=x2(t)+rand().v.(Gbest-x2(t)) (2)
Wherein: x2(t)、x3(t) indicate that second update position and third time of the individual wolf in the t times iterative process update position
It sets；GbestFor the current optimal head wolf position of wolf pack；V is the vision radius of wolf；Rand () is a random number in [- 1,1]；
Step 5.6: updating position x according to the third time that formula (3) find the individual wolf for the condition described in step 5.4 that is unsatisfactory for3
(t), wolf pack head wolf optimal location is updated, j=3 enters step 5.7；
x3(t)=x2(t)+escape().s.(xc(t)-x2(t)) (3)
Wherein, xc(t) it indicates in the t times iterative process, all individual wolves carry out second after updating, the centroid position of wolf pack；s
For the moving step length of wolf, s≤v；Escape () is random function；
Step 5.7: will be ranked up from low to high by updated all individual wolves three times according to fitness value；
Step 5.8: n+1 individual wolf for coming front is searched for the optimal position of each individual by non-linear simple method (NM method)
It sets, remaining individual repeats step 5.4- step 5.6 and searches for the optimal position of each individual wolf, obtains optimal wolf pack；From optimal wolf pack
Middle selection global optimum head wolf position；
Wherein, n is random positive integer；
Step 5.9: judging whether to reach maximum number of iterations or reach maximum search precision, if reaching, export the overall situation most
The weight and threshold value of the corresponding mobile robot electricity low frequency prediction model based on PID neural network in excellent head wolf position；Otherwise,
T+1 is enabled, step 5.3 is gone to, carries out next iteration.
4. method according to claim 1-3, which is characterized in that according to mobile robot by ground guide
The terrestrial reference with unique ID of spaced set, judges whether mobile robot breaks down, detailed process is as follows:
Mobile robot will return to remote controllers by the landmark information of ground guide in real time, when mobile robot returns
It is long-range to control if at the appointed time do not pass back through the information of next terrestrial reference a+1 in range after information by terrestrial reference a
Device judges that current mobile robot breaks down, and according to the position of terrestrial reference a and next terrestrial reference a+1, judges movement of being out of order
Position where robot.
5. according to the method described in claim 4, it is characterized in that, when mobile robot is when ground rail breaks down, far
Range controller sends track switch open command, and the track switch on ground rail is opened, and send traction instruction to break down
Mobile robot is located at the spare moving robot of same ground rail, and spare moving robot is enabled to be moved to the shifting broken down
Mobile robot position region starts the electromagnetic attraction device of spare moving robot base, adsorbs the movement broken down
Robot draws the mobile robot to break down by track switch to mobile robot region to be repaired along ground guide；
It is equipped with automatically controlled track switch between adjacent ground guide, and is provided between mobile robot region to be repaired and ground guide
Automatically controlled track switch；
The spare moving robot is located at standby machine people's waiting area, and the standby machine people waiting area is located at transport and appoints
Business starting point, and be connected to ground guide.
6. according to the method described in claim 5, it is characterized in that, when spare moving robot reaches the moving machine to break down
Device people by the last one terrestrial reference a after, open spare moving robot vehicle-mounted binocular camera ZED, measure with occur therefore
The distance between mobile robot of barrier enables spare moving robot close to the movement broken down according to measured distance
Robot；When the mobile robot for executing transport task breaks down, while being towed to maintenance area, remote controllers hair
Instruction is sent to replace failed machines people to execute transport task to the standby machine people of same track being in idle condition.
7. according to the method described in claim 6, it is characterized in that, mobile robot region to be repaired and charging guide rail connect
It is logical, if entering shifting along charging guide rail if the standby machine people's electricity being located in mobile robot region to be repaired is lower than 25%
Otherwise the charging of mobile robot charged area enters along charging guide rail positioned at the standby machine people Accreditation Waiting Area of transport task starting point
Domain.
8. a kind of intelligence manufacture environment robot and vehicle computational intelligence drive delivery system, which is characterized in that including tabletop machine
Device people, desktop control, mobile robot, guide rail and remote controllers；
The guide rail includes desktop guide rail and ground guide, and automatically controlled track switch, and ground guide are equipped between adjacent ground guide
On be arranged at intervals with the terrestrial reference with unique ID；
Automatically controlled track switch on the desktop control, mobile robot and guide rail is communicated with the remote controllers；
The desktop machine people is set on workbench, is controlled by desktop control, and the desktop guide rail on workbench moves, work
Make to be provided with the fixed object location that takes, puts on platform, and take, put object location equipped with photosensitive sensor, the photosensitive sensor with
Desktop control is connected；
Workbench is equipped with the fixed side for taking, putting object location and is provided with ranging sensing receiver；
The mobile robot is equipped with distance measuring sensor, ground is set to, using the described in any item methods of claim 1-7
Ground guide on ground moves.
9. system according to claim 8, which is characterized in that further include mobile robot charged area and standby machine people
Waiting area；
The mobile robot charged area is set on the charging guide rail between transport task initial position and end position, institute
It states charging guide rail to be connected to standby machine people's waiting area guide rail, and the standby machine people waiting area guide rail and ground guide
Connection.
10. system according to claim 9, which is characterized in that further include mobile robot region to be repaired, the movement
It is provided with automatically controlled track switch between robot region to be repaired and ground guide, and is provided with electromagnetic attraction on mobile robot pedestal
Device and vehicle-mounted binocular camera ZED, the vehicle-mounted binocular ZED are arranged in above the pedestal of robot.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810063041.3A CN108255180B (en) | 2018-01-23 | 2018-01-23 | A kind of intelligence manufacture environment robot and the vehicle computational intelligence driving means of delivery and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810063041.3A CN108255180B (en) | 2018-01-23 | 2018-01-23 | A kind of intelligence manufacture environment robot and the vehicle computational intelligence driving means of delivery and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108255180A CN108255180A (en) | 2018-07-06 |
| CN108255180B true CN108255180B (en) | 2019-05-03 |
Family
ID=62742314
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810063041.3A Active CN108255180B (en) | 2018-01-23 | 2018-01-23 | A kind of intelligence manufacture environment robot and the vehicle computational intelligence driving means of delivery and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108255180B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1103726A (en) * | 1992-11-16 | 1995-06-14 | 株式会社Pfu | Versatile production system |
| CN104992248A (en) * | 2015-07-07 | 2015-10-21 | 中山大学 | Microgrid photovoltaic power station generating capacity combined forecasting method |
| CN205798934U (en) * | 2016-07-28 | 2016-12-14 | 平湖拓伟思自动化设备有限公司 | A kind of intelligent manufacturing system based on industrial robot |
| CN107253195A (en) * | 2017-07-31 | 2017-10-17 | 中南大学 | A kind of carrying machine human arm manipulation ADAPTIVE MIXED study mapping intelligent control method and system |
| CN107272670A (en) * | 2016-04-08 | 2017-10-20 | 中国国际海运集装箱(集团)股份有限公司 | AGV system for drawing plant drying oven |
-
2018
- 2018-01-23 CN CN201810063041.3A patent/CN108255180B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1103726A (en) * | 1992-11-16 | 1995-06-14 | 株式会社Pfu | Versatile production system |
| CN104992248A (en) * | 2015-07-07 | 2015-10-21 | 中山大学 | Microgrid photovoltaic power station generating capacity combined forecasting method |
| CN107272670A (en) * | 2016-04-08 | 2017-10-20 | 中国国际海运集装箱(集团)股份有限公司 | AGV system for drawing plant drying oven |
| CN205798934U (en) * | 2016-07-28 | 2016-12-14 | 平湖拓伟思自动化设备有限公司 | A kind of intelligent manufacturing system based on industrial robot |
| CN107253195A (en) * | 2017-07-31 | 2017-10-17 | 中南大学 | A kind of carrying machine human arm manipulation ADAPTIVE MIXED study mapping intelligent control method and system |
Non-Patent Citations (4)
| Title |
|---|
| 一种机器人搬运生产线的调度优化方法及实验平台设计;张树林;《信息科技辑》;20170615;第I140-242页 |
| 基于小波神经的动力电池SOC估计的研究;于洋等;《电力电子技术》;20120630;第46卷(第6期);第90-92页 |
| 基于模糊控制的通信电源电池监测系统的设计与研究;邓颖;《工程科技II辑》;20080915;第C042-185页 |
| 复杂环境下AGVS调度系统设计;凌剑;《工程科技II辑》;20170315;第C029-503页 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108255180A (en) | 2018-07-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112799386B (en) | Robot path planning method based on artificial potential field and reinforcement learning | |
| CN108287548B (en) | A kind of automation guide rail toter and the robot collaboration means of delivery and system | |
| CN107037812A (en) | A kind of vehicle path planning method based on storage unmanned vehicle | |
| CN110963209A (en) | Garbage sorting device and method based on deep reinforcement learning | |
| Burns et al. | Anticipatory on-line planning | |
| CN106297357A (en) | Real-time route planning based on car networking and POI searching system and method | |
| CN106325284B (en) | The robot motion planning method of identification multiple target task is searched for towards man-machine collaboration | |
| CN112344945B (en) | Indoor distribution robot path planning method and system and indoor distribution robot | |
| CN107608364A (en) | A kind of intelligent robot for undercarriage on data center's physical equipment | |
| CN107992061B (en) | A kind of wisdom laboratory machine people means of delivery and system | |
| Wu et al. | A online boosting approach for traffic flow forecasting under abnormal conditions | |
| Liu et al. | Optimal robot path planning for multiple goals visiting based on tailored genetic algorithm | |
| CN109405828A (en) | Mobile robot global optimum path planning method based on LTL-A* algorithm | |
| CN110040396A (en) | Intelligent garbage bin based on big data, machine learning is made decisions on one's own System and method for | |
| CN115380293A (en) | Options selected using meta-gradient learning actions in multi-task reinforcement learning | |
| CN108280518A (en) | A kind of distributed environment robot and the vehicle mobile interchange means of delivery and system | |
| CN117234214A (en) | Automatic shuttle for stacking industrial goods | |
| CN108255180B (en) | A kind of intelligence manufacture environment robot and the vehicle computational intelligence driving means of delivery and system | |
| Zhou et al. | Route planning for unmanned aircraft based on ant colony optimization and voronoi diagram | |
| CN115574826B (en) | National park unmanned aerial vehicle patrol path optimization method based on reinforcement learning | |
| Li et al. | Vision-based obstacle avoidance algorithm for mobile robot | |
| Dai et al. | Traffic signal control using offline reinforcement learning | |
| Cong et al. | Multi-UAVs Cooperative Detection Based on Improved NSGA-II Algorithm | |
| CN110351755B (en) | Method and device for controlling nodes | |
| CN115390584A (en) | Multi-machine collaborative search method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |