CN105045094B - The task search and task executing method of multirobot group - Google Patents
The task search and task executing method of multirobot group Download PDFInfo
- Publication number
- CN105045094B CN105045094B CN201510481231.3A CN201510481231A CN105045094B CN 105045094 B CN105045094 B CN 105045094B CN 201510481231 A CN201510481231 A CN 201510481231A CN 105045094 B CN105045094 B CN 105045094B
- Authority
- CN
- China
- Prior art keywords
- robot
- task
- alliance
- cost
- search
- 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
Landscapes
- Manipulator (AREA)
Abstract
The invention discloses the task search and task executing method of a kind of multirobot group, including task search layer strategy, task search process, task receiving layer according to the different shape of task and the difference of robot cooperated form, multirobot collaboration is divided into following three kinds of forms: multirobot sequence is cooperateed with, multirobot is synchronous synergetic, multirobot freely cooperates with.Present invention can apply to execute more complicated task, by complicated Task-decomposing at several pieces, completion is cooperated by multirobot.Compared to single robot, have rapidity and high efficiency.
Description
Technical field
The present invention relates to the task search and task executing method of a kind of multirobot group.
Background technique
In recent years, it needs with the development of science and technology, robot technology is constantly progressive.Face the increasingly complicated of task
Change, single robot can no longer meet production requirement in many circumstances, and then domestic and international researcher is to multirobot skill
Art has put into a large amount of concerns, proposes using multi-robot Cooperation and single machine device is replaced manually to make.
The research of multi-robot Cooperation is not that the function by each single robot on ordinary meaning is overlapped mutually, but is led to
Each child robot cooperative cooperating in system is crossed, the effect that one-plus-one is greater than two is played.It is easily that each robot function is mutually folded
Add, not only the advantage of multi-robot system can not be highlighted, instead when multi-robot system encounters complex situations, will lead to each machine
The device human world clashes, or even stagnation or deadlock, reduces the execution efficiency of task.
Summary of the invention
The purpose of the present invention is to provide a kind of task search of multirobot group that working performance is excellent and task to hold
Row method.
The technical solution of the invention is as follows:
A kind of task search and task executing method of multirobot group, it is characterized in that: including:
(1) task search layer strategy
Assuming that the location of the robot in task search layer is initial position, the node of initial position is S, robot
From initial position, random motion search task, the search radius r=10 of robot in alliance, if the task in environment
With robot distance d at [0,10], indicate that robot searches the task；
When robot searches task, system is numbered according to the sequencing that robot searches task, to keep away
Exempt to select to clash when robot, system selects robot according to the sequencing of mission number；Since task is one quiet
State amount oneself can not select robot to execute the task, therefore place m ant in each task, use ant as dynamic
The factor replaces the suitable robot of task choosing or Federation execution task；If task representated by ant is loose type task,
Select the smallest single robot of cost i.e. executable；If the task that ant represents is close coupling type task, which need to be by more
A robot cooperates completion, is solved using corresponding coalition formation algorithm to the task, selects suitable robot
Form the Federation execution task；
Assuming that the task is close coupling type task, then need to be cooperated jointly completion by multiple robots, first put m ant
It is placed in n robot, selects suitable robot to cooperate execution task according to ant group algorithm, wherein formula (1-1) indicates
The wherein probability of kth ant selection robot j are as follows:
α is the weight of ant pheromones intensity on path, communication overhead power of the β between robot i to robot j
Weight, allowkIndicate the non-selected collection of bots of kth ant, ηij(t) communication between robot i to robot j
Overhead intensive；
If finding that current composed multirobot alliance has been able to execute when ant completes current robot selection
The task, then ant stops seeking diameter, and until m ant completes one cycle, alliance's cost is minimum in this circulation of final choice
One group of alliance, robot as current optimal solution；Pheromone update, more new formula such as formula (1-3), (1- are carried out to the overall situation simultaneously
4), (1-5):
τij(t+n)=(1- ρ) × τij(t)+Δτij(t)(1-3)
Wherein CostkIndicate the cost that robot k is paid in the task of execution；
(2) task search process
The robot task search and task distribution in task search layer are completed from following steps:
(1) robot of task Distribution Layer is from initial position, the search radius of random search task, robot
10, if task is within the scope of robot search radius, then it represents that search the task；Conversely, not having then；
(2) after robot searches task, the type of task is judged, if task is loose type task, robot transformation
The task, jump procedure (6) are executed for task execution layer；If task is close coupling type task, (3) are gone to step, to other machines
Device human hair cloth cooperative information selects suitable robot composition alliance, cooperates the task of completion；
(3) m ant is placed in each close coupling type task, task is replaced with ant, carries out probability using formula (1-1)
Selection selects the smallest robot group of cost to be combined into a new Federation execution task；
(4) it if the alliance, robot that current Ant Search arrives meets the condition for executing the task, stops search, then turns to walk
Suddenly (5)；Otherwise it (3) is gone to step continually looks for next machine and coalize；
(5) when m ant completion an iteration operation, select alliance, total cost the smallest that group of robot as optimal
Solution, and global information element is updated, as shown in formula (1-3), (1-4), (1-5)；
(6) it is finally completed search and the Robot Selection of task, executes task；
(3) task receiving layer
Task receiving layer is searched what the selected multirobot alliance after task formed by task search layer, is
The executor of task；Task receiving layer is defined as two states: task idle state, state value S=0；Execution status of task,
State value S=1；If robot is in S=0 state, indicates that task search layer can cooperate to its release tasks and request, and the machine
Device people meets the execution condition of task, then the task that executes is added in multirobot alliance, and robot goes to state 1 by state 0；If
S=1, then it represents that the robot or alliance, robot are still active, then task search layer can not be assisted to its release tasks
It requests, needs to wait for its completion task rear and be subjected to other task requests；According to the different shape of task and robot cooperated
Multirobot collaboration is divided into following three kinds of forms: the collaboration of multirobot sequence, the synchronous association of multirobot by the difference of form
Same, multirobot freely cooperates with.
The multirobot sequence collaborative strategy is based on how to select suitable robot from alliance, robot to hold
Row task is a kind of method based on cost and distance, and basic process is as follows: assuming that robot RiWhen receiving task T its
The cost consumed in unit distance is Ai, robot RiThe distance of movement is di, then robot RiConsumed by this section of path
Shown in total cost such as formula (1-7):
Costk(i)=Ai×di (1-7)
Robot consumed cost on each section of path is write down, by robot R1Start to calculate when execution task on each path
Cost summation finally obtains the total cost on each path, and is compared to it, selects in the smallest that path of cost summation
Robot composition alliance cooperate with jointly completion the task；Assuming that execute this task by m robot, then before (m-1) each machine
Shown in people's cost summation such as formula (1-8):
M-th of robot executes shown in the task cost public affairs such as formula (1-9):
Cost (m)=Am×dmT (1-9)
Total cost Cost are as follows:
Cost=Cost (m-1)+Cost (m) (1-10)
Wherein dmTFor the distance of m-th of robot to target position；Therefore, n scheme institute, alliance of n robot composition
The cost set of composition are as follows:
Cost=[Cost (1), Cost (2), Cost (3) ..., Cost (n)] selects cost summation most from set Cost
Small robot Federation execution task.
The multimachine device synchronous collaboration refers to multiple robots in multirobot alliance while being connected to multiple and different appoint
Business, each robot sets out to respectively being received for task simultaneously；When each robot reaches the task location respectively received, to each
It is analyzed from the ability of task, wherein if certain members have found that its ability is unsatisfactory for the execution condition of task, it must Xiang Lian
Other robot sends help information, plea for aid in alliance；If there is other idle machines people in the multirobot alliance, and the spare time
The ability for setting robot meets assistance condition, then idle machine people can assist the robot of scarce capacity to complete task；If alliance
In be in idle state there is no other robot member, then must wait other robot completion task, assist scarce capacity
Robot completes task；
The synchronous synergetic strategy of multirobot:
If multirobot alliance C is made of m platform robot, set are as follows: R=[R1,R2,R3,…,Rm], alliance C institute is right
The Efficiency analysis answered are as follows:The Efficiency analysis of task are as follows:, wherein there is the reception of n platform robot
It to task order, sets out to respective task, each robot assesses respective task ability when reaching task；If it was found that machine
Device people RiAbility bi<bl, then it represents that the robot capability is unsatisfactory for the condition of completion task, must seek help from other robot；If connection
There are more robots to meet assistance condition in alliance, places m ant, in the robot with robot RiLocation is
Initial position carries out optimizing, selects suitable robot to execute task using the probability mechanism of ant group algorithm, probability selection is public
Shown in formula such as formula (1-11):
If ant seeks robot RjShi Faxian has been able to satisfy the execution condition of task, then stops search；Conversely, continuing
Search；By Nc iterative search, the smallest one group of alliance, the robot assist people R of cost is selectediCompletion task.
The multirobot, which freely cooperates, refers to the search mission free in the work environment of Zhong Ge robot, alliance at random,
It not constraining by other robot, each robot belongs to task search layer at this time, and each robot is all independent individual in alliance,
It will not cause task that can not continue because some robot breaks down；
The free collaborative strategy of multirobot:
Assuming that all robot members in initial time multirobot alliance C are in free movement state, each robot
The ability radius of search mission is r；As robot RiWith task TjDistance dijWhen≤r, then it represents that robot RiSearch this
Business, robot RiTo task TjIt is assessed；If as robot R in allianceiT is taken office in search in placej, then the robot
Role transforming will be carried out, is changed into task execution layer from task search layer；Robot R simultaneouslyiTo task TjIt is analyzed, if should
Robot determines that the task for loose type task, is then independently executed by the robot；If the robot judges the task for tight coupling
Mould assembly task, then into alliance, other members issue help information, and other robot is waited to assist；System is selected according to ant group algorithm
It selects suitable robot and executes task, with cost minimum principle alternatively suitable robot Federation execution Mission Rules Guidelines；If
When C Zhong Duotai robot, alliance simultaneously scans for multiple tasks, according to robot location's distribution and the capacity of water of each robot
Selection is suitable to assist object.
Compared to traditional single robot manipulating task, multi-robot system of the present invention has following 6 advantages:
In face of more complex task, due to the limitation of its ability, application field is single for single robot, and multirobot can
It is applied to multiple fields by recombination.
Excessively single compared to single robot function, multirobot can be applied to execute more complicated task, by complexity
Task-decomposing cooperates completion at several pieces, by multirobot.Compared to single robot, have rapidity and high efficiency.
Compared to the single machine people for a certain task design complexity, cost is lower for the design of multirobot, and design is easy, is clever
Activity is good, versatile, has expanded the application field of robot.
Multi-robot system has strong stronger redundancy, robustness and concurrency.
Can be achieved to be in communication with each other between multirobot, carry out information sharing, can effectively, rapidly control the overall situation, realize fast
Speed positioning.
For multi-robot system compared with single robot system, flexibility is stronger, and system can be adopted according to task condition difference
Take different task execution schemes.
Present invention will be further explained below with reference to the attached drawings and examples.
Fig. 1 is task search layer random search task schematic diagram.
Fig. 2 is task search flow diagram.
Fig. 3 is that robot executes task path figure.
Fig. 4 is robot R under the conditions of clear1Schematic diagram is individually performed.
Fig. 5 is robot R under the conditions of clear1、R2Mutually it is performed in unison with task schematic diagram.
Fig. 6 is robot R under the conditions of clear1、R2、R3Mutually it is performed in unison with task schematic diagram.
Fig. 7 is that task schematic diagram is individually performed by robot R1 under the conditions of having obstacle.
Fig. 8 is that robot R1, R2 are mutually performed in unison with task schematic diagram under the conditions of having barrier.
Fig. 9 is that robot R1, R2, R3 are mutually performed in unison with task schematic diagram under the conditions of having barrier.
Figure 10 is the synchronous reception task schematic diagram of robot R1, R2, R3 under the conditions of clear.
Figure 11 is robot R2 assist people R3 completion task schematic diagram under the conditions of clear.
Figure 12 is robot R1, R2 assistance R3 completion task schematic diagram under the conditions of clear.
Figure 13 is the robot task selection schematic diagram in the case of having idle robot.
Figure 14 is cooperation sequential schematic of the idle machine people R4 in alliance, robot lower left.
Figure 15 is cooperation sequential schematic of the idle machine people R4 inside alliance, robot.
Figure 16 is cooperation sequential schematic of the idle machine people R4 in alliance, robot upper right side.
Figure 17 is that robot R1, R2, R3 synchronous task receives schematic diagram under the conditions of having barrier.
Figure 18 is that the robot R2 under the conditions of having barrier assists R3 to complete task schematic diagram.
Figure 19 is robot R1, R2, R3 while to receive task schematic diagram under the conditions of having barrier.
Figure 20 is that robot R4 assists R1, R3 to execute task schematic diagram under the conditions of having barrier.
Figure 21 is each robot random search task schematic diagram.
Figure 22 is that robot R2, R5 assist R1, R3 that R4 is assisted to execute task schematic diagram.
Figure 23 is that robot R5, R2, R4 assist R6 to execute task schematic diagram.
Specific embodiment
A kind of task search and task executing method of multirobot group, comprising:
(1) task search layer strategy
Assuming that the location of the robot in task search layer is initial position, the node of initial position is S, robot
From initial position, random motion search task, the search radius r=10 of robot in alliance, if the task in environment
With robot distance d at [0,10], indicate that robot searches the task；
When robot searches task, system is numbered according to the sequencing that robot searches task, to keep away
Exempt to select to clash when robot, system selects robot according to the sequencing of mission number；Since task is one quiet
State amount oneself can not select robot to execute the task, therefore place m ant in each task, use ant as dynamic
The factor replaces the suitable robot of task choosing or Federation execution task；If task representated by ant is loose type task,
Select the smallest single robot of cost i.e. executable；If the task that ant represents is close coupling type task, which need to be by more
A robot cooperates completion, is solved using corresponding coalition formation algorithm to the task, selects suitable robot
Form the Federation execution task；
Assuming that the task is close coupling type task, then need to be cooperated jointly completion by multiple robots, first put m ant
It is placed in n robot, selects suitable robot to cooperate execution task according to ant group algorithm, wherein formula (1-1) indicates
The wherein probability of kth ant selection robot j are as follows:
α is the weight of ant pheromones intensity on path, communication overhead power of the β between robot i to robot j
Weight, allowkIndicate the non-selected collection of bots of kth ant, ηij(t) communication between robot i to robot j
Overhead intensive；
If finding that current composed multirobot alliance has been able to execute when ant completes current robot selection
The task, then ant stops seeking diameter, and until m ant completes one cycle, alliance's cost is minimum in this circulation of final choice
One group of alliance, robot as current optimal solution；Pheromone update, more new formula such as formula (1-3), (1- are carried out to the overall situation simultaneously
4), (1-5):
τij(t+n)=(1- ρ) × τij(t)+Δτij(t) (1-3)
Wherein CostkIndicate the cost that robot k is paid in the task of execution；
(2) task search process
The robot task search and task distribution in task search layer are completed from following seven steps:
(1) robot of task Distribution Layer is from initial position, the search radius of random search task, robot
10, if task is within the scope of robot search radius, then it represents that search the task；Conversely, not having then；
(2) after robot searches task, the type of task is judged, if task is loose type task, robot transformation
The task, jump procedure (6) are executed for task execution layer；If task is close coupling type task, (3) are gone to step, to other machines
Device human hair cloth cooperative information selects suitable robot composition alliance, cooperates the task of completion；
(3) m ant is placed in each close coupling type task, task is replaced with ant, carries out probability using formula (1-1)
Selection selects the smallest robot group of cost to be combined into a new Federation execution task；
(4) it if the alliance, robot that current Ant Search arrives meets the condition for executing the task, stops search, then turns to walk
Suddenly (5)；Otherwise it (3) is gone to step continually looks for next machine and coalize；
(5) when m ant completion an iteration operation, select alliance, total cost the smallest that group of robot as optimal
Solution, and global information element is updated, as shown in formula (1-3), (1-4), (1-5)；
(6) it is finally completed search and the Robot Selection of task, executes task；
(3) task receiving layer
Task receiving layer is searched what the selected multirobot alliance after task formed by task search layer, is
The executor of task；Task receiving layer is defined as two states: task idle state, state value S=0；Execution status of task,
State value S=1；If robot is in S=0 state, indicates that task search layer can cooperate to its release tasks and request, and the machine
Device people meets the execution condition of task, then the task that executes is added in multirobot alliance, and robot goes to state 1 by state 0；If
S=1, then it represents that the robot or alliance, robot are still active, then task search layer can not be assisted to its release tasks
It requests, needs to wait for its completion task rear and be subjected to other task requests；According to the different shape of task and robot cooperated
Multirobot collaboration is divided into following three kinds of forms: the collaboration of multirobot sequence, the synchronous association of multirobot by the difference of form
Same, multirobot freely cooperates with.
Multirobot sequence collaboration is similar to logistics transportation, i.e., task is passed to next machine by a robot
People cooperates completion.Its premise calls is to guarantee that a certain range of task can be performed in each member in alliance, robot,
Alliance is to be executed by a robot or cooperated by more robots according to the task cost size judgement of the task of completion
To complete.The basic principle of sequence collaboration is as follows: assuming that a multirobot alliance is made of n platform robot, system, robot
The set of system are as follows: R=[R1,R2,R3,…,Rm], the cost collection that every robot is paid in unit distance in the task of execution
It is combined into: A=[A1,A2,A3,…,An].Assuming that robot R1(i.e. the first robot) receives task first, according to multirobot
Robot system known to sequence Cooperation rule has method in n to complete the task.The formulation of multirobot sequence Cooperation rule is as follows, the
I platform robot will participate in executing the task, then the execution that (i-1) the platform robot of the front must be already engaged in this task is worked as
In.
As shown in figure 3, a Federation execution task being made of three robots.First by robot R1Reception is taken office
Business, there are three types of the methods for executing the task.The first scheme, by robot R1The task is individually performed, it is straight via path d14
It connects and arrives at the destination.Second scheme, by robot R1With robot R2Common cooperation is completed, via the d12 in path by path
D24 is arrived at the destination, and completes task.The third scheme is by robot R1, robot R2And robot R3Mutually collaboration is completed,
It is finally arrived at the destination by path d34 by path d1 by path d23.How to select robot to execute the strategy of task is machine
One of the research emphasis of device people cooperation.
The multirobot sequence collaborative strategy is based on how to select suitable robot from alliance, robot to hold
Row task is a kind of method based on cost and distance, and basic process is as follows: assuming that robot RiWhen receiving task T its
The cost consumed in unit distance is Ai, robot RiThe distance of movement is di, then robot RiConsumed by this section of path
Shown in total cost such as formula (1-7):
Costk(i)=Ai×di (1-7)
Robot consumed cost on each section of path is write down, by robot R1Start to calculate when execution task on each path
Cost summation finally obtains the total cost on each path, and is compared to it, selects in the smallest that path of cost summation
Robot composition alliance cooperate with jointly completion the task；Assuming that execute this task by m robot, then before (m-1) each machine
Shown in people's cost summation such as formula (1-8):
M-th of robot executes shown in the task cost public affairs such as formula (1-9):
Cost (m)=Am×dmT (1-9)
Total cost Cost are as follows:
Cost=Cost (m-1)+Cost (m) (1-10)
Wherein dmTFor the distance of m-th of robot to target position；Therefore, n scheme institute, alliance of n robot composition
The cost set of composition are as follows:
Cost=[Cost (1), Cost (2), Cost (3) ..., Cost (n)] selects cost summation most from set Cost
Small robot Federation execution task.
Emulation experiment and analysis
(1) the multirobot sequence under clear environment cooperates with
When executing the task of same position in different location to robot in multirobot alliance respectively herein (Task),
Collaboration selection between robot is studied.Assuming that a multirobot alliance is by robot R1、R2、R3Composition, system to
First mission bit stream gives robot R to alliance's publication1, alliance holds according to the cost size selection task that robot executes the task
Row scheme.By table 1-1 it is found that respectively indicating robot R1、R2、R3Different schemes is selected to execute this at different positions
Business.
Multirobot sequence collaboration costing analysis under the conditions of table 1-1 clear
From table 1-1: when Zhong Ge robot, alliance is distributed in R1(10,50), R2(60,30), R3When (40,60), by machine
Device people R1It is minimum that task cost is individually performed；When Zhong Ge robot, alliance is distributed in R1(15,65), R2(40,40), R3(70,
20), then by robot R1, R2Cooperate cost minimum；When Zhong Ge robot, alliance is distributed in R1(10,80), R2(30,50),
R3(50,40), then by robot R1, R2, R3Cost minimum is cooperated, simulation result is as shown in Figure 4,5, 6.It can thus be appreciated that
When robot location changes in alliance, the strategy that robot executes task can also change therewith.
(2) the multirobot sequence under obstacle environment cooperates with
Multirobot work compound and the Path selection under clear environment are had studied above, and this section mainly studies machine
How people alliance selects robot Federation execution task in the environment of with the presence of barrier.Robot is passed in the task of execution
Sensor detects robot far from barrier distance d < dsaveWhen, then robot rotates to the right θ angle by rule, and continues to the right
Move to safe distance S1, move to robot after safe distance and rotate clockwise 90 ° and continue to move S2Return to the position originally moved
It sets, formula are as follows: S2=S1× tan θ, robot obstacle-avoiding as shown in figs. 1-7, horizontal direction indicate S1, vertical direction expression S2。
By table 1-2 when robot is in R1(30,50), R2(80,40), R3When (80,30) this three positions it is found that by
Robot R1Total cost is minimum when task is individually performed, and Path selection is as shown in figs. 1-7.Although robot R1To R2、R3Between
Clear interference, and robot R1There are barrier interference between target point, it has been investigated that robot is by getting around barrier
The cost for hindering object finally to consume is less than and R2、R3The cost paid when cooperating, therefore alliance's selection is by robot R1Individually hold
Row task.As three robot R1(40,70), R2(60,50), R3(90,30) be in three positions when, by table 1-2 it is found that
By robot R1、R2The cost that cooperates is minimum, and Path selection is as shown in Figure 8.And when three robots are in following three position R1
(40,70), R2(70,40), R3When (50,15), by table 1-2 it is found that by robot R1, R2, R3Cost minimum is completed in cooperation, is imitated
True result is as shown in Figure 9.Simulation result shows even if, there are barrier, robot also can be by keeping away between robot and task
It opens barrier to cooperate, the final optimal solution for obtaining problem.
In conclusion even if there are barrier between robot and task, robot can be executed by getting around barrier
Task, the cost that the cost consumed expends when may also cooperate than the robot with other clears is small, table 1-2
Demonstrate the robot cooperation relation when there is barrier.
Table 1-2 has the multirobot sequence collaboration costing analysis under the conditions of barrier
The multimachine device synchronous collaboration refers to multiple robots in multirobot alliance while being connected to multiple and different appoint
Business, each robot sets out to respectively being received for task simultaneously；When each robot reaches the task location respectively received, to each
It is analyzed from the ability of task, wherein if certain members have found that its ability is unsatisfactory for the execution condition of task, it must Xiang Lian
Other robot sends help information, plea for aid in alliance；If there is other idle machines people in the multirobot alliance, and the spare time
The ability for setting robot meets assistance condition, then idle machine people can assist the robot of scarce capacity to complete task；If alliance
In be in idle state there is no other robot member, then must wait other robot completion task, assist scarce capacity
Robot completes task；
The synchronous synergetic strategy of multirobot:
If multirobot alliance C is made of m platform robot, set are as follows: R=[R1,R2,R3,…,Rm], alliance C institute is right
The Efficiency analysis answered are as follows:The Efficiency analysis of task are as follows:, wherein there is the reception of n platform robot
It to task order, sets out to respective task, each robot assesses respective task ability when reaching task；If it was found that machine
Device people RiAbility bi<bl, then it represents that the robot capability is unsatisfactory for the condition of completion task, must seek help from other robot；If connection
There are more robots to meet assistance condition in alliance, places m ant, in the robot with robot RiLocation is
Initial position carries out optimizing, selects suitable robot to execute task using the probability mechanism of ant group algorithm, probability selection is public
Shown in formula such as formula (1-11):
If ant seeks robot RjShi Faxian has been able to satisfy the execution condition of task, then stops search；Conversely, continuing
Search；By Nc iterative search, the smallest one group of alliance, the robot assist people Ri of cost is selected to complete task.
Emulation experiment and analysis
(1) multirobot under clear environment is synchronous synergetic
When the member in alliance simultaneously receive system publication task order, set out to respective task, such as Figure 10 institute
Show, wherein cross symbol indicates the ability that the robot has execution task, triangle symbol indicates that the robot does not have
The ability of execution task, demand help other members in alliance.The robot R known to table 1-31And R2All has assist people R3
The ability of completion task carries out probability selection according to ant group algorithm, and as shown in table 1-4, selects robot R2Assist people R3
Cost is optimal, and simulation result is as shown in figure 11.
Table 1-3 robot capability and task ability
1-4 assist people R3Execution task cost
If working as robot R2It assists not to be able to satisfy condition still in participating in, then must continue searching other robot and participate in association
It helps, each robot capability is as shown in tables 1 to 5.
Each robot capability of table 1-5 and task ability
By table 1-5 it is found that the ability value of task 3 is 80, only robot R in set1、R2.Therefore, by robot R1With
R2Common assist people R3Task could be completed, simulation result is as shown in figure 12.
If other robot, in the task of execution, there are idle machine people in alliance, and the robot capability is able to satisfy association
Condition is helped, then the robot is preferentially selected to assist completion task.Robot R in alliance as shown in figure 131And R3It cannot complete to appoint
Business, needs robot R4It assists to complete, wherein cross symbol indicates that the robot can complete task, triangle symbol indicates the machine
Device people cannot complete task.
(2) there is the multirobot under obstacle environment synchronous synergetic
Multirobot in the case where there is obstacle environment is synchronous synergetic, when executing task when encountering barrier motion mode with
Mode is identical in sequence collaboration.By current environment Xia Yousantai robot known to table 1-7, each robot capability and each is indicated in table
Task ability.As shown in Figure 17, robot R1, R2, R3 receives task T1, T2, T3 simultaneously, and sets out to task, robot
Barrier is encountered in the way of R1 execution task, robot senses barrier by sensor in advance, in advance avoiding obstacles.Figure
In cross symbol indicate the robot have execute the task ability, triangle symbol indicate the robot do not have execution should
The ability of task.By table 1-7 it is found that robot R3 needs to wait for other robot completion task assists its completion again.Machine in table 1-7
Device people R1 and R2, which have, assists the task ability, according to minimum cost principle, as shown in table 1-8, and final choice robot R2
Assist people R3 execution is optimal, and simulation result is as shown in figure 18.
Table 1-7 has each robot capability under the conditions of barrier
Table 1-8 has the assist people R under the conditions of barrier3Execution task cost
As shown in figure 19, robot R1, R3Ability be unsatisfactory for the condition of execution task, while existing in alliance, robot
One idle machine people R4, and R4Meet assist people R1, R3The ability of completion task, using the TSP problem of ant group algorithm as mould
Type solves shortest path, final robot R4First assist people R1Execution task, then assist R3Execution task, simulation result
As shown in figure 20.
The multirobot, which freely cooperates, refers to the search mission free in the work environment of Zhong Ge robot, alliance at random,
It not constraining by other robot, each robot belongs to task search layer at this time, and each robot is all independent individual in alliance,
It will not cause task that can not continue because some robot breaks down；
The free collaborative strategy of multirobot:
Assuming that all robot members in initial time multirobot alliance C are in free movement state, each robot
The ability radius of search mission is r；As robot RiWith task TjDistance dijWhen≤r, then it represents that robot RiSearch this
Business, robot RiTo task TjIt is assessed；If as robot R in allianceiT is taken office in search in placej, then the robot
Role transforming will be carried out, is changed into task execution layer from task search layer；Robot R simultaneouslyiTo task TjIt is analyzed, if should
Robot determines that the task for loose type task, is then independently executed by the robot；If the robot judges the task for tight coupling
Mould assembly task, then into alliance, other members issue help information, and other robot is waited to assist；System is selected according to ant group algorithm
It selects suitable robot and executes task, with cost minimum principle alternatively suitable robot Federation execution Mission Rules Guidelines；If
When C Zhong Duotai robot, alliance simultaneously scans for multiple tasks, according to robot location's distribution and the capacity of water of each robot
Selection is suitable to assist object.
Emulation experiment and analysis
Herein by taking transfer robot as an example, robot search mission random in 100 × 100 place, robot
The ability radius of search mission is 10.The position distribution after task is searched for each robot as shown in figure 21, wherein cross
Symbol indicates that the robot searches task, and triangle indicates that robot does not seek task.Wherein table 1-9 indicates each in alliance C
Robot capability, table 1-10 indicate the ability of each task.
Robot capability value in table 1-9 alliance
The ability value of each task of 1-10
As shown in Figure 21, robot R1、R4、R6Task T is searched respectively1、T2、T3, due to robot R6With task away from
Within the task search radius in robot, robot R6Task has just been searched in initial position.By table 1-9
With shown in table 1-10, robot R6Ability be unsatisfactory for the requirement of execution task, need into alliance other free movements machine
People requests help, and according to minimum cost principle, using the probability selection mechanism of ant group algorithm, final choice is by robot R2、R5Association
Help robot R1, robot R3Assist R4Execution task is the optimal solution of problem, and simulation result is as shown in figure 22.
Each robot of table 1-11 executes the cost of task to task 3
Since other robot is in execution status of task, R in alliance6Other robot need to be waited to complete task in situ
Afterwards, then its completion is assisted.When other robot completion task, robot appropriate is selected to execute task, as shown in table 1-11, when
By robot R5、R2、R4Cooperate robot R6Cost is minimum, but robot R2、R3、R4Ability be not able to satisfy execution task
Condition, by comparison, finally by robot R5、R2、R4Assist R6Execution task is optimal, simulation result such as Figure 23 of problem
It is shown.
Claims (1)
1. a kind of task search and task executing method of multirobot group, it is characterized in that: including:
(1) task search layer strategy
Assuming that the location of robot in task search layer is initial position, the node of initial position is S, and robot is from first
Beginning position is set out, random motion search task, the search radius r=10 of robot in alliance, if task and machine in environment
Device people's distance d indicates that robot searches the task at [0,10]；
When robot searches task, system is numbered according to the sequencing that robot searches task, to avoid selecting
It is clashed when selecting robot, system selects robot according to the sequencing of mission number；Since task is a static amount,
Oneself robot can not be selected to execute the task, therefore place m ant in each task, use ant as dynamic factor generation
For the suitable robot of task choosing or Federation execution task；If task representated by ant is loose type task, generation is selected
The smallest single robot of valence is i.e. executable；If the task that ant represents is close coupling type task, which need to be by multiple machines
People cooperates completion, is solved using corresponding coalition formation algorithm to the task, selects suitable robot composition connection
Alliance executes the task；
Assuming that the task is close coupling type task, then need to be cooperated jointly completion by multiple robots, first be placed in m ant
In n robot, suitable robot is selected to cooperate execution task according to ant group algorithm, wherein formula (1-1) indicates wherein
The probability of kth ant selection robot j are as follows:
Weight of the α for ant pheromones intensity on path, communication overhead weight of the β between robot i to robot j,
allowkIndicate the non-selected collection of bots of kth ant, ηij(t) communication overhead between robot i to robot j
Intensity；dijFor robot RiWith task TjDistance；
If finding that current composed multirobot alliance has been able to execute this when ant completes current robot selection
Business, then ant stops seeking diameter, and until m ant completes one cycle, alliance's cost is the smallest by one in this circulation of final choice
Group alliance, robot is as current optimal solution；Simultaneously to the overall situation carry out Pheromone update, more new formula such as formula (1-3), (1-4),
(1-5):
τij(t+n)=(1- ρ) × τij(t)+Δτij(t) (1-3)
Wherein CostkIndicate the cost that robot k is paid in the task of execution；
(2) task search process
The robot task search and task distribution in task search layer are completed from following six step:
(1) robot of task Distribution Layer is from initial position, and random search task, the search radius of robot is 10, if
Task is within the scope of robot search radius, then it represents that searches the task；Conversely, not having then；
(2) after robot searches task, judge the type of task, if task is loose type task, robot, which is changed into, to appoint
Business execution level executes the task, jump procedure (6)；If task is close coupling type task, (3) are gone to step, to other robot
Cooperative information is issued, suitable robot composition alliance is selected, cooperates the task of completion；
(3) m ant is placed in each close coupling type task, task is replaced with ant, carries out probability choosing using formula (1-1)
It selects, the smallest robot group of cost is selected to be combined into a new Federation execution task；
(4) it if the alliance, robot that current Ant Search arrives meets the condition for executing the task, stops search, then goes to step
(5)；Otherwise it (3) is gone to step continually looks for next machine and coalize；
(5) when m ant completion an iteration operation, select alliance, total cost the smallest that group of robot as optimal solution, and
Global information element is updated, as shown in formula (1-3), (1-4), (1-5)；
(6) it is finally completed search and the Robot Selection of task, executes task；
(3) task receiving layer
Task receiving layer is searched what the selected multirobot alliance after task formed by task search layer, is task
Executor；Task receiving layer is defined as two states: task idle state, state value S=0；Execution status of task, state
Value S=1；If robot is in S=0 state, indicates that task search layer can cooperate to its release tasks and request, and the robot
The execution condition for meeting task, then be added in multirobot alliance the task that executes, and robot goes to state 1 by state 0；If S=
1, then it represents that the robot or alliance, robot are still active, then task search layer can not cooperate to its release tasks and ask
It asks, needs to wait for its completion task rear and be subjected to other task requests；According to the different shape of task and robot cooperated form
Difference, multirobot collaboration is divided into following three kinds of forms: the collaboration of multirobot sequence, multirobot are synchronous synergetic, more
Robot freely cooperates with；
Multirobot sequence cooperates with, i.e., task is passed to next robot by a robot, cooperate completion；Before it
Making demands is to guarantee that a certain range of task can be performed in each member in alliance, robot, and alliance appoints according to the task of completion
Business cost size judgement is to be executed by a robot or cooperate to complete by more robots；Sequence cooperates with basic
Principle is as follows: assuming that a multirobot alliance is made of n platform robot, the set of robot system are as follows: R=[R1,R2,
R3,Λ,Rm], the cost set that every robot is paid in unit distance in the task of execution are as follows: A=[A1,A2,A3,Λ,
An]；Assuming that robot R1, i.e., the first robot receives task, according to multirobot sequence Cooperation rule machine first
People's system has method in n to complete the task；The formulation of multirobot sequence Cooperation rule is as follows, and i-th robot will participate in executing
The task, then (i-1) the platform robot of the front must be already engaged in the execution of this task；
When a Federation execution task being made of three robots, first by robot R1Task is received, there are three types of hold
The method of the row task；The first scheme, by robot R1The task is individually performed, goes directly to destination via path d14；
Second scheme, by robot R1With robot R2Common cooperation is completed, and reaches purpose by path d24 via the d12 in path
Task is completed on ground；The third scheme is by robot R1, robot R2And robot R3Mutually collaboration is completed, and is passed through by path d1
Path d23 is finally arrived at the destination by path d34；
The multirobot sequence collaborative strategy is based on how that suitable robot is selected to appoint to execute from alliance, robot
Business, is a kind of method based on cost and distance, basic process is as follows: assuming that robot RiIts unit when receiving task T
Cost apart from upper consumption is Ai, robot RiThe distance of movement is di, then robot RiConsumed total generation on this section of path
Shown in valence such as formula (1-7):
Costk(i)=Ai×di (1-7)
Robot consumed cost on each section of path is write down, by robot R1Start to calculate when execution task on each path
Cost summation finally obtains the total cost on each path, and is compared to it, selects in the smallest that path of cost summation
Robot composition alliance cooperate with jointly completion the task；Assuming that execute this task by m robot, then before (m-1) each machine
Shown in people's cost summation such as formula (1-8):
M-th of robot executes shown in the task cost public affairs such as formula (1-9):
Cost (m)=Am×dmT (1-9)
Total cost Cost are as follows:
Cost=Cost (m-1)+Cost (m) (1-10)
Wherein dmTFor the distance of m-th of robot to target position；Therefore, n scheme of alliance of n robot composition is formed
Cost set are as follows:
Cost=[Cost (1), Cost (2), Cost (3), Λ, Cost (n)] selects cost summation the smallest from set Cost
The robot Federation execution task；
The synchronous synergetic multiple robots referred in multirobot alliance of the multimachine device while it being connected to multiple and different tasks, respectively
Robot sets out to respectively being received for task simultaneously；When each robot reaches the task location respectively received, appoint to respective
The ability of business is analyzed, wherein if certain members have found that its ability is unsatisfactory for the execution condition of task, it must be into alliance
Other robot sends help information, plea for aid；If when there are other idle machines people in multirobot alliance, and idle machine
The ability of device people meets assistance condition, then idle machine people can assist the robot of scarce capacity to complete task；If in alliance not
There are other robot members to be in idle state, then other robot must be waited to complete task, assist the machine of scarce capacity
People completes task；
The synchronous synergetic strategy of multirobot:
If multirobot alliance C is made of m platform robot, set are as follows: R=[R1,R2,R3,Λ,Rm], energy corresponding to alliance C
Force vector are as follows:Wherein there is n platform robot to receive task order, goes out to respective task
Hair, each robot assesses respective task ability when reaching task；If it was found that robot RiAbility bi<bl, then it represents that it should
Robot capability is unsatisfactory for the condition of completion task, must seek help from other robot；If there are more robots to meet in alliance
Assistance condition places m ant, in the robot with robot RiLocation is that initial position carries out optimizing, is utilized
The probability mechanism of ant group algorithm selects suitable robot to execute task, shown in probability selection formula such as formula (1-11):
If ant seeks robot RjShi Faxian has been able to satisfy the execution condition of task, then stops search；Conversely, continuing searching；
By Nc iterative search, the smallest one group of alliance, the robot assist people Ri of cost is selected to complete task；
The multirobot freely cooperates with the search mission free in the work environment for referring to that Zhong Ge robot, alliance is random, not by
Other robot constraint, each robot belongs to task search layer at this time, and each robot is all independent individual in alliance, will not
Cause task that can not continue because some robot breaks down；
The free collaborative strategy of multirobot:
Assuming that all robot members in initial time multirobot alliance C are in free movement state, each robot search
The ability radius of task is r；As robot RiWith task TjDistance dijWhen≤r, then it represents that robot RiThe task is searched,
Robot RiTo task TjIt is assessed；If as robot R in allianceiT is taken office in search in placej, then the robot will be real
Row role transforming is changed into task execution layer from task search layer；Robot R simultaneouslyiTo task TjIt is analyzed, if the machine
People determines that the task for loose type task, is then independently executed by the robot；If the robot judges the task for close coupling type
Task, then into alliance, other members issue help information, and other robot is waited to assist；System selects to close according to ant group algorithm
Suitable robot executes task, with cost minimum principle alternatively suitable robot Federation execution Mission Rules Guidelines；If alliance C
When Zhong Duotai robot simultaneously scans for multiple tasks, select to close according to the capacity of water of robot location's distribution and each robot
Suitable assistance object.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510481231.3A CN105045094B (en) | 2015-08-03 | 2015-08-03 | The task search and task executing method of multirobot group |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510481231.3A CN105045094B (en) | 2015-08-03 | 2015-08-03 | The task search and task executing method of multirobot group |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105045094A CN105045094A (en) | 2015-11-11 |
| CN105045094B true CN105045094B (en) | 2019-05-17 |
Family
ID=54451714
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510481231.3A Active CN105045094B (en) | 2015-08-03 | 2015-08-03 | The task search and task executing method of multirobot group |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105045094B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TWI812923B (en) * | 2021-02-09 | 2023-08-21 | 法博智能移動股份有限公司 | Dynamic configuration method based on role assignation, electronic device, and medium |
Families Citing this family (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106341483B (en) * | 2016-09-28 | 2019-09-10 | 深圳市普渡科技有限公司 | A kind of distributed scheduling system communication means based on LAN multicast group |
| CN106529776B (en) * | 2016-10-31 | 2017-11-24 | 合肥工业大学 | The autonomous cotasking distribution method of multiple agent |
| CN106737696A (en) * | 2017-03-11 | 2017-05-31 | 杭州极木科技有限公司 | Multi-robot cooperative transportation method |
| US11025566B2 (en) * | 2017-04-12 | 2021-06-01 | Facebook, Inc. | Techniques for intent-based search for bots |
| CN107560631B (en) * | 2017-08-30 | 2020-02-14 | 国网智能科技股份有限公司 | Path planning method and device and inspection robot |
| CN107831685B (en) * | 2017-10-13 | 2023-03-14 | 南方科技大学 | Group robot control method and system |
| CN108460488A (en) * | 2018-03-12 | 2018-08-28 | 舟谱数据技术南京有限公司 | Deep learning Intelligent logistics Allocation plan system based on prediction perception |
| CN110309993B (en) * | 2018-03-27 | 2022-04-05 | 杭州海康机器人技术有限公司 | AGV task allocation method and device for automatic guided transport vehicle |
| CN108803598A (en) * | 2018-05-08 | 2018-11-13 | 南方科技大学 | A kind of polar region multiple robots operating system and Synergistic method |
| CN110046795B (en) * | 2019-03-01 | 2021-11-05 | 斯坦德机器人（深圳）有限公司 | Task allocation method and device for robot |
| CN110083166A (en) * | 2019-05-30 | 2019-08-02 | 浙江远传信息技术股份有限公司 | For the coordinated dispatching method of multirobot, device, equipment and medium |
| CN110347159B (en) * | 2019-07-12 | 2022-03-08 | 苏州融萃特种机器人有限公司 | Mobile robot multi-machine cooperation method and system |
| CN111080258B (en) * | 2019-12-18 | 2020-11-17 | 中国人民解放军军事科学院国防科技创新研究院 | Group unmanned system cooperative task management subsystem based on role state machine |
| CN112060097B (en) * | 2020-11-10 | 2021-02-05 | 纳博特南京科技有限公司 | Multitask cooperative control method of multi-robot system |
| CN112434779B (en) * | 2020-12-09 | 2023-11-21 | 哈尔滨工程大学 | Improved ant colony algorithm-based tight coupling task allocation method |
| CN114721839B (en) * | 2022-06-09 | 2022-09-06 | 中国人民解放军国防科技大学 | Method and device for detecting and optimizing deadlock abnormal data of robot group task allocation |
| CN115860431B (en) * | 2023-02-07 | 2023-05-26 | 广东技术师范大学 | Heterogeneous perception-based intelligent scheduling method, system, robot and medium for multiple robots |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101364110A (en) * | 2008-09-28 | 2009-02-11 | 重庆邮电大学 | Cooperating work control method and system for robot of multiple degree of freedom |
| CN102831319A (en) * | 2012-08-27 | 2012-12-19 | 北京科技大学 | Task allocation algorithm of fine-grained multi-robot collaboration system |
| CN104317293A (en) * | 2014-09-19 | 2015-01-28 | 南京邮电大学 | City rescue intelligent agent dynamic path planning method based on improved ant colony algorithm |
-
2015
- 2015-08-03 CN CN201510481231.3A patent/CN105045094B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101364110A (en) * | 2008-09-28 | 2009-02-11 | 重庆邮电大学 | Cooperating work control method and system for robot of multiple degree of freedom |
| CN102831319A (en) * | 2012-08-27 | 2012-12-19 | 北京科技大学 | Task allocation algorithm of fine-grained multi-robot collaboration system |
| CN104317293A (en) * | 2014-09-19 | 2015-01-28 | 南京邮电大学 | City rescue intelligent agent dynamic path planning method based on improved ant colony algorithm |
Non-Patent Citations (4)
| Title |
|---|
| automated synthesis of multi-robot task solution through software reconfiguration;Fang Tang，等;《Proceeding of the 2005 IEEE International Conference on Robotics and Automation Barcelona, Spain》;20050430;第1501-1508页 * |
| Multi-robot Dynamic Task Allocation Using Modified Ant Colony System;Xu Z，等;《International Conference on Artificial Intelligence and Computational Intelligence. Springer-Verlag》;20091231;第5855卷;第288-297页 * |
| 基于群体智能的多机器人任务分配;刘淑华，等;《吉林大学学报(工学版)》;20100131;第40卷(第1期);摘要、第123-129页及图1 * |
| 多机器人系统协同作业技术发展近况与前景;周东健，等;《机电技术》;20131231(第6期);第146-150页 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TWI812923B (en) * | 2021-02-09 | 2023-08-21 | 法博智能移動股份有限公司 | Dynamic configuration method based on role assignation, electronic device, and medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105045094A (en) | 2015-11-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105045094B (en) | The task search and task executing method of multirobot group | |
| Wu et al. | Modeling and deadlock control of automated guided vehicle systems | |
| CN101945492B (en) | Clustering-based multi-robot task allocation method | |
| Kruse et al. | Exploiting human cooperation in human-centered robot navigation | |
| CN110497403A (en) | A kind of manipulator motion planning method improving two-way RRT algorithm | |
| CN102023571B (en) | Clustering-based multi-robot task distributing method for use in exploiting tasks | |
| CN110083166A (en) | For the coordinated dispatching method of multirobot, device, equipment and medium | |
| CN106979785A (en) | A kind of complete traverse path planing method of multi-robot system | |
| Liu et al. | Path planning based on ant colony algorithm and distributed local navigation for multi-robot systems | |
| CN111273664B (en) | Multi-robot path collaborative planning method for space truss structure on-orbit assembly | |
| CN101618543A (en) | Task allocation method of heterogeneous multi-robot system | |
| CN110375761A (en) | Automatic driving vehicle paths planning method based on enhancing ant colony optimization algorithm | |
| CN109947120A (en) | Paths planning method in warehousing system | |
| CN111522341A (en) | Multi-time-varying formation tracking control method and system for network heterogeneous robot system | |
| CN110231821A (en) | The adaptive kernel action amalgamation method of the improvement of multi-robot formation | |
| JP2009199359A (en) | Autonomous robot and cooperative working system using a plurality of autonomous robots | |
| Nie et al. | Research on robot path planning based on Dijkstra and Ant colony optimization | |
| CN112327890A (en) | Underwater multi-robot path planning based on WHCA algorithm | |
| CN104834809A (en) | Artificial colony search-based seven-degree-of-freedom mechanical arm reverse kinematical solving method | |
| CN113189988A (en) | Autonomous path planning method based on Harris algorithm and RRT algorithm composition | |
| Zhang et al. | Learning to cooperate: Application of deep reinforcement learning for online AGV path finding | |
| Takimoto et al. | Saving energy consumption of multi-robots using higher-order mobile agents | |
| Guo et al. | Distributed task allocation algorithm based on connected dominating set for WSANs | |
| Juntao et al. | Study on robot path collision avoidance planning based on the improved ant colony algorithm | |
| Abe et al. | Searching targets using mobile agents in a large scale multi-robot environment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |