CN111240319B - Outdoor multi-robot cooperative operation system and method thereof - Google Patents
Outdoor multi-robot cooperative operation system and method thereof Download PDFInfo
- Publication number
- CN111240319B CN111240319B CN201911415926.6A CN201911415926A CN111240319B CN 111240319 B CN111240319 B CN 111240319B CN 201911415926 A CN201911415926 A CN 201911415926A CN 111240319 B CN111240319 B CN 111240319B
- Authority
- CN
- China
- Prior art keywords
- robot
- task
- path planning
- master
- scheme
- 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
Images
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course or altitude of land, water, air, or space vehicles, e.g. automatic pilot
- 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
- G05D1/0234—Control of position or course in two dimensions specially adapted to land vehicles using optical position detecting means using optical markers or beacons
- G05D1/0236—Control of position or course in two dimensions specially adapted to land vehicles using optical position detecting means using optical markers or beacons in combination with a laser
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course or altitude of land, water, air, or space vehicles, e.g. automatic pilot
- 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
- G05D1/0214—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory in accordance with safety or protection criteria, e.g. avoiding hazardous areas
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course or altitude of land, water, air, or space vehicles, e.g. automatic pilot
- 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 or altitude of land, water, air, or space vehicles, e.g. automatic pilot
- 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
- G05D1/0223—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory involving speed control of the vehicle
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course or altitude of land, water, air, or space vehicles, e.g. automatic pilot
- 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
- G05D1/0238—Control of position or course in two dimensions specially adapted to land vehicles using optical position detecting means using obstacle or wall sensors
- G05D1/024—Control of position or course in two dimensions specially adapted to land vehicles using optical position detecting means using obstacle or wall sensors in combination with a laser
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course or altitude of land, water, air, or space vehicles, e.g. automatic pilot
- 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
- G05D1/0255—Control of position or course in two dimensions specially adapted to land vehicles using acoustic signals, e.g. ultra-sonic singals
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course or altitude of land, water, air, or space vehicles, e.g. automatic pilot
- 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
- G05D1/0257—Control of position or course in two dimensions specially adapted to land vehicles using a radar
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course or altitude of land, water, air, or space vehicles, e.g. automatic pilot
- 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
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/02—Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS]
Abstract
The invention discloses an outdoor multi-robot cooperative operation system and a method thereof, wherein the system comprises a master control background and a mobile robot group, the master control background comprises a communication module, a display device and a server, the mobile robot group is composed of a master robot and a plurality of slave robots based on a master-slave structure, and each robot is provided with a navigation positioning module, a motion control module, a sensor and a communication and master control module.
Description
Technical Field
The invention belongs to the field of intelligent robots, and particularly relates to an outdoor multi-robot cooperative operation system.
Background
With the rapid development of the robot technology, the robot field tends to develop towards the direction of clustering and collaborative operation. With the increasing complexity of the operation task, the capability of a single robot in the aspects of information acquisition, real-time processing, operation range and the like is very limited, and the operation requirement is difficult to achieve in some occasions; in addition, it may be easier and more economical to perform a task with multiple inexpensive robots than it is with a more intelligent, more expensive, single robot. Due to the advantage that multiple robots cooperatively complete tasks, people pay more and more attention to the robot, and the robot is more and more widely applied to various fields, such as post-disaster search and rescue, automatic production lines, ocean and space exploration, military and the like.
A multi-robot cooperative operation system relates to a multi-robot path planning technology, a multi-robot communication technology, a multi-robot cooperative control technology and other key technologies, and the research difficulty of the existing multi-robot cooperative operation system is mainly concentrated on the aspects of path planning and task allocation. The conventional multi-robot collaborative path planning method mainly comprises the following steps: artificial potential field method, distributed control method, particle swarm algorithm, coevolution gene algorithm method, etc. The Artificial Potential Field Method is excellent in real-time control, but the Potential Field function is complex to construct, solving the problem of extreme values requiring solving partial differential equations with finite difference Method, computationally intensive to solve, complex, and prone to trapping in local minima (Tian-tie ZHAO. Path Planning of Mobile Robot Based on Bacterial formation and Engineering Point Field Method [ C ]. Science and Engineering Research Center. Proceedings of20172 and national Conference Test, measurementation and Computational Method (TMCM 2017.) Science and Engineering Research Center: science and Engineering Research Center, 2017; the distributed control method has high requirements on robot individuals, and although the distributed control method has excellent flexibility and robustness, the manufacturing cost of the mobile robot group is high (Wukuping. Research on global navigation technology of mobile robots based on distributed vision [ D ]. University of Hebei industry, 2007.); particle swarm algorithm computation of Multi-constrained composite Path Planning problem is fast, but also tends to fall into local optimality (Li Di. Robot Path Planning Based on Improved Multi-Objective PSO Method [ C ]. Wuhan horizontal time Development Co., proceedings of the International Conference Computer engineering, information science and application Technology (ICCIA 2016.) Wuhan horizontal time Development Co., 2016-513); solving a multi-robot collaborative Path Planning problem Based on a Genetic Algorithm of collaborative evolution, with the increase of constraints, the calculation time is greatly increased, the solving speed is slow, and the real-time performance is poor (Gan xushing. Robot Path Planning Based on Genetic and chastic Optimization Algorithm [ C ]. Research Institute of Management Science and Industrial Engineering of 20175th International Conference on Computer, automation and Power Electronics (cap 2017.) Research Institute of Management Science and Industrial Engineering: computer Science and Electronic Technology International Society, 2017-120.
At present, the task allocation of multi-robot cooperative work mainly comprises two categories, namely a centralized type and a distributed type. The centralized task allocation algorithm can effectively solve the problems of task conflict and saturated working conditions when the number of the robots is small, but as the dimension and the number of the mobile robots are increased, the information amount and the communication traffic of the robots are exponentially increased, and the calculation with high complexity is difficult to adapt (a multi-robot task allocation method combining pre-allocation with Hungarian algorithm, CN109615188A, 2019.04.12); distributed algorithms are less affected by the number of robots and have better adaptability and robustness, but a 'saturation effect' exists, and data transfer among non-co-located modules generates inter-robot communication, so that increasing the number of robots can cause the reduction of the throughput of the whole system on the contrary (Qin Xin Li et al, multi-robot task allocation [ J ] space control technology and application based on the improved ant colony algorithm, 2018,44 (05): 55-59.).
Disclosure of Invention
The invention aims to provide an efficient and reliable outdoor multi-robot cooperative operation system and a method thereof.
The technical solution for realizing the purpose of the invention is as follows: an outdoor multi-robot cooperative operation system comprises a master control background and a mobile robot group, wherein the master control background comprises a communication device, a display device and a server; the mobile robot group is based on a master-slave structure and comprises a master robot and at least one slave robot, wherein the master robot and the slave robot are provided with navigation positioning, motion control, sensors, communication and master control modules.
An outdoor multi-robot cooperative operation method comprises the following steps:
firstly, an environment map is built, after the mobile robot group finishes putting, a master control background remote control host robot detects environment information, and a grid environment map is built through high-precision laser radar scanning;
marking task points, and marking the task points of different grades on the constructed grid environment map by the master control background to obtain coordinates of the task points;
thirdly, distributing tasks, and distributing the tasks based on graph theory and genetic algorithm according to the Tth i Task point coordinates (X) oi ，Y oi ) Level L and M i Coordinates (x) of mobile robot group i ，y i ) Firstly, obtaining an iteration initial value by using a KM method based on a bipartite graph, and then iterating through a genetic algorithm to obtain a minimum cost scheme;
and fourthly, planning a path, namely planning sub-layer paths by a single mobile robot according to the result of task allocation in the previous step, planning a main-layer path by a master control background to obtain a total minimum cost path scheme, and performing cooperative operation by the master robot and the slave robot according to the total minimum cost path scheme.
Compared with the prior art, the invention has the advantages that: (1) The mobile robot group adopts a master-slave structure, only the master robot is provided with the high-precision laser radar environment perception sensor, and the overall manufacturing cost of the robot is low; (2) The master robot of the mobile robot group has a slave robot backup, and a collision avoidance strategy is adopted for coordination processing in the operation process, so that the reliability and the robustness of the system are improved; (3) The invention adopts a task allocation algorithm based on graph theory and genetic algorithm, firstly obtains an iteration initial value by using a weighted bipartite graph, and then iterates a final scheme by combining the genetic algorithm, and the time complexity of the task allocation algorithm is low; (4) The invention adopts a path planning algorithm based on a master-slave structure, the sub-layer planning obtains a single robot shortest path scheme by simulating an annealing ant colony algorithm, the main layer planning uses a differential evolution algorithm to comprehensively consider the positions of all robots and the single robot shortest path scheme, and the overall minimum cost scheme of the mobile robot group is obtained by adjustment, so that the problem of local optimization is effectively avoided, the resolving speed is high, and the real-time performance is high.
The present invention will be described in further detail with reference to the accompanying drawings.
Drawings
FIG. 1 is a schematic diagram of information interaction between a general control background and mobile robot population according to the present invention.
FIG. 2 is a schematic diagram of information interaction between the general control background and the mobile robot group and between the general control background and the mobile robot.
Fig. 3 is a flowchart of the work of the outdoor multi-robot cooperative work system.
Fig. 4 is a schematic diagram of a path planning algorithm based on a master-slave structure in the present invention.
Fig. 5 is a flow chart of the operation of the group of mobile robots of the present invention.
Detailed Description
With reference to fig. 1, the outdoor multi-robot cooperative operation system of the present invention includes a master control background and a mobile robot group composed of multiple robots, wherein the master control background includes a communication module, a display device and a server, and monitors the running state of the mobile robot group in real time; the mobile robot group is based on a master-slave structure and comprises a master robot and at least one slave robot, wherein the master robot and the slave robot are respectively provided with a navigation positioning module, a motion control module, a sensor, a communication module and a master control module.
The main robot sends environment map information to the master control background through the high-precision laser radar and the communication module; the mobile robot group sends real-time position information to the master control background through the navigation positioning module and the communication module; the mobile machine crowd sends the sub-layer path planning scheme to a master control background through a communication module; the master control background monitors the running state of the mobile robot in real time, and when the mobile robot is out of the working state, a locking command is sent; and the master control background comprehensively considers the positions of the robots and the sub-layer path planning scheme of the single robot, carries out master layer path planning adjustment to obtain a total minimum cost scheme, and sends the master layer path planning scheme to the mobile robot through the communication module.
With reference to fig. 2, there are four slave robots. The motion control module is composed of 1 motion control panel, 4 straight motors and 4 steering motors, wherein the motion control panel issues instructions to control the running states of the 8 motors through a canopen protocol, receives obstacle avoidance sensor data through a 485 protocol, and receives main control module instructions and uploads sensor data through an Ethernet protocol.
The communication modules of the master robot and the slave robot are the same and are composed of a wireless AP and two antennas, the mobile robot group and the master control background are in networking communication through the wireless AP, and the master control background communication module and each robot are provided with the wireless AP and 2 antennas.
The sensor module mainly includes keeps away barrier sensor, temperature and humidity sensor and dropproof sensor, keeps away the barrier sensor and keeps away the barrier sensor for 8 probe ultrasonic waves, and the dropproof sensor is 2 probe dropproof ultrasonic sensor, sends detected data for motion control panel through 485 agreements in real time.
The navigation positioning module is a laser radar environment perception sensor and a cartographer algorithm software function package, wherein the slave robot is provided with a common laser radar environment perception sensor, and the master robot is provided with a high-precision laser radar environment perception sensor (in practice during laser scanning, an included angle between two beams of laser is not more than 0.25 degrees, the high precision is high, and the angle is more than 0.25 degrees, the common is adopted);
the main control module is an industrial personal computer, receives sensor data uploaded by the motion control panel and sending instructions through an Ethernet protocol, and receives laser radar environment sensor data through the Ethernet protocol, so that the functions of constructing an environment map, navigating and positioning and sub-layer path planning are realized.
The main robot completes the following work:
scanning surrounding environment information by using a high-precision laser radar environment perception sensor, receiving laser radar data by a main control module, constructing a work scene grid map by using a cartographer algorithm, and sending the work scene grid map to a main control background server through a communication module;
real-time position information is sent to a master control background through a navigation positioning module and a communication module of a main robot;
the main control module of the main robot is used for processing the pose information returned by the navigation positioning module, planning a sub-layer path scheme from the main robot to a task point and sending the sub-layer path scheme to a main control background server;
the main control module receives a main layer path planning scheme planned by the main control background and executes tasks according to instructions of the main control background;
the main control module of the main robot is used for processing the returned information of the obstacle avoidance sensor, so that the obstacle avoidance and obstacle crossing functions in the operation process are realized;
the slave robot performs the following tasks:
receiving a working scene map sent by a master control background by using a communication module of the slave robot;
real-time position information is sent to a master control background through a navigation positioning module and a communication module of the slave robot;
processing information returned by the navigation positioning module by using a master control module of the slave robot, planning a sub-layer path planning scheme of the master control module, and sending the sub-layer path planning scheme to a master control background server;
the main control module receives a main layer path planning scheme planned by the master control background and executes tasks according to instructions;
the master control module of the slave robot is used for processing the information returned by the obstacle avoidance sensor, so that the obstacle avoidance and obstacle crossing functions in the operation process are realized;
the master control background completes the following work:
the communication module is used for exchanging data with the mobile robot group in real time, receiving environment map information constructed by the main robot and receiving position information of the mobile robot group;
setting the level of the task point as a first-level task point and a second-level task point, and marking the task point on an environment map;
the server adjusts and combines a total optimization scheme through a main layer path planning algorithm according to a sub-layer path planning scheme of the mobile machine crowd, and issues the scheme to the mobile machine crowd through a communication module;
when the mobile robot group can not finish the task independently, the mobile robot group is remotely controlled;
and displaying the running state of the mobile robot in real time, judging whether a fault exists or not and whether the fault is separated from the working state or not, and if so, giving out sound and light alarm.
Except the information interaction between the master control background and the mobile robots, the information interaction between the mobile robots is as follows: after the mobile robot group finishes putting, the master robot starts to construct a working environment map and sends the working environment map to the slave robot through the communication module; after the slave robot is started, waiting for receiving an environment map constructed by the master robot; when two robots in the mobile robot group enter an obstacle avoidance range, an obstacle avoidance coordination strategy is started to be executed until no other mobile robots are in the effective range of the obstacle avoidance sensor.
The master robot is provided with a high-precision laser radar environment perception sensor, one high-precision laser radar environment perception sensor which is configured as same as the master robot is selected from the slave robots to be used as a backup, and when the master robot cannot work normally, the slave robot receives a master control background instruction to become the master robot.
The task point position is the coordinate where the mobile robot group works, the mobile robot group works is calibrated, the task point level is a first-level task point and a second-level task point, the first-level task point is a target which can be independently worked and completed by a single mobile robot, and the second-level task point is a target which needs two or more mobile robots to cooperatively work and complete.
With reference to fig. 3, the method for the outdoor multi-robot cooperative operation of the present invention comprises the following steps:
the method comprises the steps of firstly, constructing an environment map, after a mobile robot group is thrown into a working environment, detecting environment information by a master control background remote control host robot, and constructing a grid environment map through high-precision laser radar scanning.
Marking task points, and marking the task points of different grades on the constructed grid environment map by the master control background to obtain coordinates of the task points; the task point position is the coordinate of a point where a mobile robot group works, the task point level is a first-level task point and a second-level task point, the first-level task point is a target which can be independently completed by a single mobile robot, and the second-level task point is a target which needs two or more mobile robots to cooperatively complete.
Thirdly, task allocation is carried out, a task allocation algorithm based on graph theory and genetic algorithm is adopted on a master control background server, and the task allocation algorithm is carried out according to the Tth i Task point coordinates (X) oi ，Y oi ) Level L and M i Coordinates (x) of mobile robot group i ，y i ) Firstly, obtaining an iteration initial value by using a KM method based on a bipartite graph, and then iterating through a genetic algorithm to obtain a minimum cost scheme;
and fourthly, path planning, namely performing sub-layer path planning by a single mobile robot master control module according to the result of task allocation in the previous step, and performing master layer path planning by a master control background server to obtain a total minimum cost path scheme, wherein the master robot and the slave robot perform cooperative operation according to the total minimum cost path scheme.
After the environment map is constructed, the mobile robot group obtains self pose information by the following steps:
the first step, particle initialization, randomly setting N sample points, w is equal to the weight of each sample 0 i =1/N, obtaining an initial set of particles
Second, the particles are sampled, using the motion model of the robot as a proposed distribution in which the set S of particles from the previous generation is collected k-1 And generating a new particle set, namely performing pose estimation on each particle through a motion model. At this time
Thirdly, calculating weight value by using the observed value z of the laser radar k For the sample set S k Updating the weight of each particle in the distribution list, wherein the weight is the ratio of the actual distribution to the proposed distribution
May be derived from a likelihood model of the sensor.
And fourthly, resampling is carried out according to the weight of the particles, the particles with large weight represent higher possibility, and the resampling is increased. Weight of the resampled particles
And fifthly, estimating the pose, namely calculating the mean value and the variance of the particle set according to the particle set at the moment k, wherein the mean value is the pose estimation of the robot.
The invention relates to a task allocation algorithm based on graph theory and genetic algorithm, which comprises the following steps:
the method comprises the steps of firstly, according to an optimization target with minimum sum weighting of a mobile robot work load balancing function F (x) and a total moving distance function d (x), establishing a minimum cost scheme objective function F (x) = min (omega) 1 f(x)+ω 2 d(x))，ω 1 +ω 2 ＝1。
Second, load task point T i Coordinate (X) of oi ，Y oi ) Complexity information O i And a mobile robot M i Coordinate (x) of i ，y i ) Execution capability A i And so on.
Thirdly, establishing a weighted bipartite graph from the mobile robot group M to the task point T, and setting the weight value as the robot M i To task point T i Assigning a value to each vertex, assigning the vertex of the mobile machine group M as the maximum weight of the edge connected with the vertex, and assigning the vertex of the task point T as 0.
And fourthly, starting matching, wherein the matching principle is that only the edges with the same weight as the M vertexes are matched, and if the edge matching cannot be found, the assignment of all the M vertexes of the path is subtracted by d, and the assignment of all the T vertexes is added by d.
And fifthly, matching again, if each T vertex is matched with the corresponding M vertex, finishing the iteration initial value, and otherwise, returning to the fourth step.
Sixthly, initializing a genetic algorithm cross mutation mode and probability P i Parameters such as population size S and iteration number N.
Step seven, initializing a population of the genetic algorithm to form a feasible task allocation scheme that each T vertex corresponds to L M vertices; task allocation needs to satisfy task point level L = number of mobile robots.
Eighthly, calculating the fitness values of all individuals in the population according to the objective function established in the first step; and selecting individuals with higher fitness in the population to form a next generation population according to the fitness value by adopting a roulette mode.
Ninthly, performing crossing, mutation and copying operations on the population after the selection operation; wherein the crossing and mutation modes adopt two-point crossing and one-point mutation respectively.
Tenth, judging whether the current iteration number is larger than a preset iteration number N, wherein N is generally less than or equal to 10000, if yes, finishing the iteration, and selecting an optimal individual from the current new generation of population, namely the individual with the minimum cost function value, as an optimal task allocation scheme of the mobile robot; otherwise, returning to execute the eighth step.
With reference to fig. 4, in the specific path planning of the present invention, the path planning of the main robot adopts a differential evolution algorithm, and if the fitness value of the new individual is better in the algorithm, the newly generated individual will replace the original individual and be matched with the function realized by the backup robot; and (4) finding an optimal solution through iteration by adopting a simulated annealing ant colony algorithm from the robot path planning.
The sub-layer path planning is a path optimization scheme that mobile machine groups plan to task points respectively, the respective path optimization scheme is planned by each mobile machine group in the sub-layer, then the respective path optimization scheme is sent to a main layer general monitoring background, and the positions of each mobile machine group and each task point are comprehensively considered on the basis of the sub-layer planning to obtain a main layer path planning scheme, namely a total minimum cost scheme. The path planning of the main robot adopts a differential evolution algorithm, and if the fitness value of a new individual is better in the algorithm, the newly generated individual replaces the original individual and is matched with the function realized by the backup robot; and (4) finding an optimal solution through iteration by adopting a simulated annealing ant colony algorithm from the robot path planning.
The sub-layer path planning method comprises the following specific steps:
firstly, initializing the number m of ants, a task point N, a cooling coefficient a, a belonging to (0, 1);
second, annealing initiation temperature setting T = T max Setting the starting points of all ants as v 0 ；
Third, for each ant k (k =1, 2...., m), j ∈ allowK indicates that the next node of ant k must be an unvisited node, according to the formula
Selecting the next node, and after the ant is transferred from node i to node j, the pheromone tau on the opposite side (i, j) ij And (6) updating. Wherein tau is 0 Is constant and xi is an adjustable parameter. According to the formula
τ ij ＝(1-ξ)τ ij +ξτ 0
Partially updating pheromones, and repeating the steps until all nodes are visited;
and fourthly, all ants finish the searching task, and the shortest distance d is updated, and the minimum passing node number n is obtained. According to the formula
Calculating the deviation of each ant to obtain the current optimal ant A 0 ；
Fifthly, the current optimal ant A is used 0 As initial solution, generating new solution based on simulated annealing principle, if the new solution is not inferior, updating d, n, and recalculating A 0 Judging whether to accept the new solution according to the acceptance probability p;
the sixth step, according to the formula
τ ij (t+1)＝(1-ρ)τ ij (t)+ρΔτ ij best ,ρ∈(0,1)
For optimal ant A 0 Carrying out global pheromone updating;
the seventh step, according to the formula, the temperature reduction formula is
T＝T·a
And (5) cooling. If T is less than T min Quit cycle output optimal ant A 0 Otherwise, go to the third step.
The method for planning the path entering the main layer comprises the following steps:
firstly, initializing an initial value of a differential evolution algorithm;
and secondly, the master layer planning realizes the adjustment and optimization of the paths of the slave robots. The individuals of the differential evolution algorithm are represented as a set of n pointers, each pointing to a path from the robot. After the individuals of the differential evolution algorithm are successfully expressed, substituting parameters of the particle swarm algorithm obtained by sub-layer path planning, and forming an adaptive function by running time and pause time through intersection and variation to obtain an optimal scheme of the current combined path;
thirdly, if the iteration times of the differential evolution algorithm are larger than the maximum iteration times, exiting the iteration; otherwise, executing the next iteration until the optimal path of the path planning module is obtained;
and fourthly, sending the optimal path planning obtained by the master control station to each robot to realize a master-slave structure multi-robot path planning task.
The paths obtained by path planning are coordinated through a collision avoidance strategy, so that the safe operation of the mobile robot group is ensured. When the safe distance is smaller than the minimum safe distance during the operation of a plurality of robots, starting a collision avoidance strategy, and adopting different measures to process different collision conditions:
first, the direction of travel of the robot is not changed by collision. If free grids exist above and below the collision grid, a robot is randomly selected to avoid the free grid, and the robot passes the free grid and then follows the original route; otherwise, adopting a backspacing strategy, withdrawing and avoiding the robot close to the evasive grid, and driving the other robot on the original line;
and secondly, side collision, wherein the running direction of the robot forms an included angle of 90 degrees with the original running direction or the included angle of the running directions of the two original robots is 90 degrees. The robot close to the task point waits for a step length of time on the collision grid, and the robot far from the task point passes through the collision grid.
With reference to fig. 5, the specific work flow of the mobile robot group of the present invention is as follows:
after the mobile robot group is started, each robot sends starting success information to the master control background;
the master control background remote control main robot constructs a working scene map through a high-precision laser radar and determines the coordinates of a target point at the same time;
the slave robot waits for receiving the map information, and the master robot sends the constructed map information to the master control background and the slave robot;
the mobile machine crowd receives the task distribution information of the master control background;
the sub-layer of the mobile machine population plans a minimum cost scheme to a task point and sends the scheme to a master control background;
the master control background receives the sub-layer planning scheme, adjusts and optimizes the sub-layer planning scheme to obtain a total minimum cost path planning scheme, and sends the scheme to the mobile machine crowd;
and the mobile robot group operates according to the received general control background instruction.
Claims (7)
1. An outdoor multi-robot cooperative operation system is characterized by comprising a master control background and a mobile robot group, wherein the master control background comprises a communication module, a display device and a server; the mobile robot group is based on a master-slave structure and consists of a master robot and at least one slave robot, wherein the master robot and the slave robots are provided with a navigation positioning module, a motion control module, a sensor, a communication module and a master control module;
the motion control module consists of a motion control board, a straight motor and a steering motor, wherein the motion control board issues instructions to control the running states of the straight motor and the steering motor and receives obstacle avoidance sensor data, main control module instructions and uploading sensor data;
each communication module consists of a wireless AP and an antenna, and the mobile robot group and the master control background are in networking communication through the wireless AP;
the sensor module comprises an obstacle avoidance sensor, a temperature and humidity sensor and a falling prevention sensor, and transmits detection data to the motion control panel in real time;
the navigation positioning module is a laser radar environment perception sensor and a cartographer algorithm software functional package, wherein the slave robot is provided with a common laser radar environment perception sensor, and the master robot is provided with a high-precision laser radar environment perception sensor;
the main control module is an industrial personal computer, receives sensor data and sending instructions uploaded by the motion control panel, receives laser radar environment sensor data, and realizes the functions of constructing an environment map, navigating and positioning and sub-layer path planning;
in the main robot:
the main robot senses and scans surrounding environment information by using a high-precision laser radar environment sensing sensor, the main control module receives laser radar data, constructs a work scene grid map by using a cartographer algorithm, and sends the work scene grid map to a main control background server through the communication module;
real-time position information is sent to a master control background through a navigation positioning module and a communication module of a main robot;
the main control module of the main robot is used for processing the pose information returned by the navigation positioning module, planning a sub-layer path scheme from the main robot to a task point and sending the sub-layer path scheme to a main control background server;
the main control module receives a main layer path planning scheme planned by the main control background and executes tasks according to instructions of the main control background;
and the main control module of the main robot is used for processing the returned information of the obstacle avoidance sensor, so that the obstacle avoidance and obstacle crossing functions in the operation process are realized.
2. The system of claim 1, wherein said slave robots:
receiving a working scene map sent by a master control background by using a communication module of the slave robot;
real-time position information is sent to a master control background through a navigation positioning module and a communication module of the slave robot;
processing information returned by the navigation positioning module by using a master control module of the slave robot, planning a sub-layer path planning scheme of the master control module, and sending the sub-layer path planning scheme to a master control background server;
the main control module receives a main layer path planning scheme planned by the master control background and executes tasks according to instructions;
and the master control module of the slave robot is used for processing the returned information of the obstacle avoidance sensor, so that the obstacle avoidance and obstacle crossing functions in the operation process are realized.
3. The system according to claim 1, wherein in the general control background:
the communication module is used for exchanging data with the mobile robot group in real time, and receiving environment map information constructed by the main robot and position information of the mobile robot group;
setting the level of the task point as a first-level task point and a second-level task point, and marking the task point on an environment map;
the server adjusts and combines a total optimization scheme through a main layer path planning algorithm according to the sub-layer path planning scheme of the mobile machine crowd, and sends the scheme to the mobile machine crowd through the communication module;
when the mobile robot group can not finish the task independently, the mobile robot group is remotely controlled;
and displaying the running state of the mobile robot in real time, judging whether a fault exists or not and whether the fault is separated from the working state or not, and if so, giving out sound and light alarm.
4. The system of claim 1, wherein the master robot has a high-precision lidar environment sensing sensor, one of the slave robots configured with the same high-precision lidar environment sensing sensor as the master robot is selected as a backup, and when the master robot cannot work normally, the slave robot receives a master control background command to become the master robot;
the task point position is the coordinate where the mobile robot group works, the mobile robot group works is calibrated, the task point level is a first-level task point and a second-level task point, the first-level task point is a target which can be independently worked and completed by a single mobile robot, and the second-level task point is a target which needs two or more mobile robots to cooperatively work and complete.
5. An outdoor multi-robot cooperative work method for the outdoor multi-robot cooperative work system of claim 1, characterized by the steps of:
firstly, an environment map is built, after a mobile robot group is thrown into a working environment, a master control background remotely controls a host robot to detect environment information, and a grid environment map is built through high-precision laser radar scanning;
marking task points, and marking the task points of different grades on the constructed grid environment map by the master control background to obtain coordinates of the task points;
thirdly, task allocation is carried out, a task allocation algorithm based on graph theory and genetic algorithm is adopted on a master control background server, and the task allocation algorithm is carried out according to the Tth i Task point coordinates (X) oi ，Y oi ) Level L and M i Coordinates (x) of mobile robot group i ，y i ) Firstly, obtaining an iteration initial value by using a KM method based on a bipartite graph, and then iterating through a genetic algorithm to obtain a minimum cost scheme;
fourthly, path planning, namely, according to the result of task allocation in the last step, a single mobile robot main control module carries out sub-layer path planning, then a master control background server carries out main-layer path planning to obtain a total minimum cost path scheme, the master robot and the slave robot carry out cooperative operation according to the total minimum cost path scheme,
after the environment map is built, the mobile robot group obtains self pose information by the following steps:
the first step, particle initialization, randomly setting N sample points, w is equal to the weight of each sample 0 i =1/N, obtaining an initial set of particles
Second, the particles are sampled, using the motion model of the robot as a proposed distribution in which the set S of particles of the previous generation is collected k-1 Generating a new set of particles, i.e. performing pose estimation on each particle through the motion model, at the moment
Thirdly, calculating weight value by using the observed value z of the laser radar k For the sample set S k The weight of each particle in (1) is updated, and the weight is the ratio of the actual distribution to the proposed distribution
Can be obtained by a likelihood model of the sensor;
and fourthly, resampling is carried out according to the weight of the particles, the particles with large weight represent higher possibility, the weight of the particles after resampling can be increased during resampling, and the weight of the particles after resampling is carried out
And fifthly, estimating the pose, namely calculating the mean value and the variance of the particle set according to the particle set at the moment k, wherein the mean value is the pose estimation of the robot.
The task allocation algorithm based on the graph theory and the genetic algorithm comprises the following steps:
the method comprises the steps of firstly, according to an optimization target with minimum sum weighting of a mobile robot work load balancing function F (x) and a total moving distance function d (x), establishing a minimum cost scheme objective function F (x) = min (omega) 1 f(x)+ω 2 d(x))，ω 1 +ω 2 ＝1；
Second, load task point T i Coordinate (X) of oi ，Y oi ) Complexity information O i And a mobile robot M i Coordinate (x) of (2) i ，y i ) Execution capability A i Attribute information is equal;
thirdly, establishing a weighted bipartite graph of the mobile robot group M to the task point T, and setting the weight as the robot M i To task point T i Assigning a value to each vertex, assigning the vertex of the mobile machine group M as the maximum weight of the side connected with the vertex, and assigning the vertex of the task point T as 0;
fourthly, starting matching, wherein the matching principle is that only the edges with the same weight as the M vertexes are matched, if the edge matching cannot be found, d is subtracted from the assignments of all the M vertexes of the path, and d is added to the assignments of all the T vertexes;
fifthly, matching again, if each T vertex is matched with the corresponding M vertex, the iteration initial value is finished, otherwise, returning to the fourth step;
sixthly, initializing a genetic algorithm cross mutation mode and probability P i Parameters such as population size S and iteration number N;
step seven, initializing a population of a genetic algorithm to form a feasible task allocation scheme that each T vertex corresponds to L M vertexes, wherein the task allocation needs to meet the task point level L = the number of mobile robots;
eighthly, calculating the fitness values of all individuals in the population according to the objective function established in the first step, and selecting the individuals with higher fitness in the population according to the fitness values in a roulette mode to form a next generation population;
performing crossing, variation and copying operations on the population after the selection operation, wherein the crossing and variation modes adopt double-point crossing and single-point variation respectively;
and tenth step, judging whether the current iteration number is larger than a preset iteration number N, wherein N is less than or equal to 10000, if so, ending the iteration, selecting the optimal individual from the current new generation of population, namely the individual with the minimum cost function value, as the optimal task allocation scheme of the mobile robot, and otherwise, returning to the eighth step.
6. The method according to claim 5, wherein in the mobile robot path planning method based on the master-slave structure, the sub-layer path planning is a path optimization scheme that the mobile robot groups plan to task points respectively, the respective path optimization scheme is planned by the respective mobile robot groups of the sub-layer, and then is sent to the main layer main monitoring background, and the positions of the mobile robots and the positions of the task points are comprehensively considered on the basis of the sub-layer path planning to obtain a main layer path planning scheme, namely a total minimum cost scheme;
the path planning of the main robot adopts a differential evolution algorithm, and if the fitness value of a new individual is better in the algorithm, the newly generated individual replaces the original individual and is matched with the function realized by the backup robot; the optimal solution is found through iteration by adopting a simulated annealing ant colony algorithm in the robot path planning;
the sub-layer path planning algorithm comprises the following specific steps:
firstly, initializing the number m of ants, a task point N, a cooling coefficient a, a belonging to (0, 1);
second, annealing initial temperature setting T = T max Setting the starting points of all ants as v 0 ；
Third, for each ant k (k =1, 2...., m), j ∈ allowK indicates that the next node of ant k must be an unvisited node, according to the formula
Selecting the next node, and after the ant is transferred from node i to node j, the pheromone tau on the opposite side (i, j) ij Performing an update wherein 0 Is constant and xi is an adjustable parameter according to the formula tau ij ＝(1-ξ)τ ij +ξτ 0 Locally updating pheromones, and repeating the steps until all nodes are visited;
fourthly, all ants finish the searching task, the shortest distance d is updated, the number n of the nodes which pass through the shortest distance is updated, and the searching task is carried out according to the formula w 1 +w 2 ＝1
Calculating the deviation of each ant to obtain the current optimal ant A 0 ；
Fifthly, the current optimal ant A is used 0 As initial solution, generating new solution based on simulated annealing principle, if the new solution is not inferior, updating d, n, and recalculating A 0 Is judged to be based on the reception probability pIf not, accepting a new solution;
the sixth step, according to the formula
τ ij (t+1)＝(1-ρ)τ ij (t)+ρΔτ ij best ,ρ∈(0,1)
For optimal ant A 0 Carrying out global pheromone updating;
seventhly, cooling according to a formula of
T＝T·a
Cooling operation if T is less than T min Quit circulation output optimal ant A 0 Otherwise, turning to the third step;
entering a main layer path planning algorithm:
firstly, initializing an initial value of a differential evolution algorithm;
secondly, the main layer planning realizes the adjustment and optimization of the paths of the slave robots, the individuals of the differential evolution algorithm are represented as a set with n pointers, each pointer points to one path of the slave robots, after the individuals of the differential evolution algorithm are successfully represented, parameters of the particle swarm algorithm obtained by the sub-layer path planning are substituted, and the optimal scheme of the current combined path is obtained by crossing and variation and then forming an adaptive function by running time and pause time; step two, if the iteration times of the differential evolution algorithm are larger than the maximum iteration times, exiting the iteration; otherwise, executing the next iteration until the optimal path of the path planning module is obtained;
and thirdly, the optimal path planning obtained by the master control station is sent to each robot, and a master-slave structure multi-robot path planning task is realized.
7. The method according to claim 5, wherein the path obtained by path planning is coordinated by a collision avoidance strategy to ensure safe operation of the mobile robot group, when the safe distance is less than the minimum safe distance during operation of a plurality of robots, the collision avoidance strategy is activated, and different collision situations are handled by different measures:
firstly, the running direction of the robot is not changed during collision, if free grids exist above and below a collision grid, the robot is randomly selected to leave the free grids for avoiding, and the robot is carried out along the original route after another robot passes through the free grids; otherwise, adopting a backspacing strategy, withdrawing and avoiding the robot close to the evaluable grid, and driving the other robot on the original line;
and secondly, side collision, wherein the traveling direction of the robot forms an included angle of 90 degrees with the original traveling direction or the included angle of the traveling directions of the two original robots is 90 degrees, the robot close to the task point waits for a step length of time on a collision grid, and the robot far away from the task point passes through the collision grid first.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911415926.6A CN111240319B (en) | 2019-12-31 | 2019-12-31 | Outdoor multi-robot cooperative operation system and method thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911415926.6A CN111240319B (en) | 2019-12-31 | 2019-12-31 | Outdoor multi-robot cooperative operation system and method thereof |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111240319A CN111240319A (en) | 2020-06-05 |
| CN111240319B true CN111240319B (en) | 2023-03-28 |
Family
ID=70872413
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911415926.6A Active CN111240319B (en) | 2019-12-31 | 2019-12-31 | Outdoor multi-robot cooperative operation system and method thereof |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111240319B (en) |
Families Citing this family (20)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111885550B (en) * | 2020-06-06 | 2023-09-19 | 浙江中力机械股份有限公司 | Distributed autonomous mobile robot scheduling system |
| CN111617415A (en) * | 2020-06-10 | 2020-09-04 | 中国船舶重工集团应急预警与救援装备股份有限公司 | Fully-autonomous fire-fighting robot operating system in complex environment and working method |
| CN111786987B (en) * | 2020-06-29 | 2023-04-25 | 杭州海康机器人股份有限公司 | Task issuing method, device, system and equipment |
| CN111708369B (en) * | 2020-07-17 | 2021-07-23 | 武汉科技大学 | Route planning method for transformer substation inspection robot |
| CN112207815B (en) * | 2020-08-21 | 2022-04-12 | 北京卫星制造厂有限公司 | Multi-machine multi-process space-time collaborative planning method and system |
| CN112232711B (en) * | 2020-11-11 | 2022-10-21 | 北京理工大学 | Multi-isomorphic unmanned vehicle distributed task allocation method for reconnaissance tasks |
| CN112731923B (en) * | 2020-12-17 | 2023-10-03 | 武汉万集光电技术有限公司 | Cluster robot co-positioning system and method |
| CN113093732B (en) * | 2021-03-12 | 2023-02-24 | 武汉联一合立技术有限公司 | Intelligent killing system and method |
| CN112883526B (en) * | 2021-03-15 | 2023-04-07 | 广西师范大学 | Workload distribution method under task delay and reliability constraint |
| CN112947489B (en) * | 2021-04-08 | 2022-11-22 | 华东理工大学 | Method and device for planning collision-free path of welding robot in complex environment |
| CN114578772B (en) * | 2021-04-16 | 2023-08-11 | 青岛中车四方车辆物流有限公司 | AGV cluster control system design framework and method |
| CN113467477B (en) * | 2021-08-03 | 2023-03-10 | 艾乐尔机器人科技(南京)有限公司 | Many intelligent robot underground cable maintenance device based on video identification technique |
| CN113916231A (en) * | 2021-09-28 | 2022-01-11 | 上海浦东发展银行股份有限公司 | Mobile equipment positioning navigation method and device, mobile equipment and storage medium |
| CN114254902B (en) * | 2021-12-13 | 2024-04-02 | 四川启睿克科技有限公司 | Multi-production-line personnel scheduling method |
| CN114879674A (en) * | 2022-05-12 | 2022-08-09 | 安徽大学绿色产业创新研究院 | Routing inspection robot scheduling and path planning method in prefabricated part storage yard |
| CN115113626B (en) * | 2022-07-20 | 2023-04-11 | 威康(深圳)智能有限公司 | Robot distributed control system and method thereof |
| CN116822826A (en) * | 2023-03-13 | 2023-09-29 | 广东博嘉拓建筑科技有限公司 | Multi-machine collaborative construction path planning method, medium and construction robot |
| CN117381802B (en) * | 2023-12-12 | 2024-03-05 | 吉林省吉邦自动化科技有限公司 | Distributed multi-robot cooperative control method |
| CN117437382B (en) * | 2023-12-19 | 2024-03-19 | 成都电科星拓科技有限公司 | Updating method and system for data center component |
| CN117726153B (en) * | 2024-02-18 | 2024-04-19 | 中国科学院工程热物理研究所 | Real-time re-planning method suitable for unmanned aerial vehicle cluster operation tasks |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108776474B (en) * | 2018-05-24 | 2022-03-15 | 中山赛伯坦智能科技有限公司 | Robot embedded computing terminal integrating high-precision navigation positioning and deep learning |
| CN109597418B (en) * | 2019-02-27 | 2021-03-02 | 福州大学 | Robot distributed cooperative obstacle avoidance method based on independent virtual center point |
| CN109986563B (en) * | 2019-05-01 | 2022-03-15 | 湖南大学 | Multi-mobile-robot cooperative operation method and system |
| CN110244772B (en) * | 2019-06-18 | 2021-12-03 | 中国科学院上海微系统与信息技术研究所 | Navigation following system and navigation following control method of mobile robot |
| 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 |
-
2019
- 2019-12-31 CN CN201911415926.6A patent/CN111240319B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN111240319A (en) | 2020-06-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111240319B (en) | Outdoor multi-robot cooperative operation system and method thereof | |
| Hu et al. | Voronoi-based multi-robot autonomous exploration in unknown environments via deep reinforcement learning | |
| Fu et al. | A heuristic evolutionary algorithm of UAV path planning | |
| CN106406338B (en) | Autonomous navigation device and method of omnidirectional mobile robot based on laser range finder | |
| CN113110457B (en) | Autonomous coverage inspection method for intelligent robot in indoor complex dynamic environment | |
| CN110989352B (en) | Group robot collaborative search method based on Monte Carlo tree search algorithm | |
| CN110471426A (en) | Unmanned intelligent vehicle automatic Collision Avoidance method based on quantum wolf pack algorithm | |
| CN112068588A (en) | Unmanned aerial vehicle trajectory generation method based on flight corridor and Bezier curve | |
| Yu et al. | Autonomous obstacle avoidance for uav based on fusion of radar and monocular camera | |
| Zhao et al. | A path planning method based on multi-objective cauchy mutation cat swarm optimization algorithm for navigation system of intelligent patrol car | |
| CN112612290B (en) | Underwater vehicle three-dimensional multi-task path planning method considering ocean currents | |
| CN112327939B (en) | Collaborative path planning method for high-rise fire-fighting multiple unmanned aerial vehicles in city block environment | |
| Vashisth et al. | Existing path planning techniques in unmanned aerial vehicles (UAVs): A systematic review | |
| Ni et al. | An improved real-time path planning method based on dragonfly algorithm for heterogeneous multi-robot system | |
| CN114186859B (en) | Multi-machine cooperative multi-target task allocation method in complex unknown environment | |
| CN112857370A (en) | Robot map-free navigation method based on time sequence information modeling | |
| Xin et al. | Coordinated motion planning of multiple robots in multi-point dynamic aggregation task | |
| Zhicai et al. | A cooperative search algorithm based on improved particle swarm optimization decision for UAV swarm | |
| CN114063614B (en) | Multi-small robot system mixed formation rapid convergence method based on airborne monocular vision perception | |
| Yong et al. | An autonomous navigation strategy based on improved hector slam with dynamic weighted a* algorithm | |
| Belavadi et al. | Frontier exploration technique for 3d autonomous slam using k-means based divisive clustering | |
| CN110989350B (en) | Method and device for realizing pose optimization of underground mobile robot based on membrane calculation | |
| Chen et al. | A real-time multi-constraints obstacle avoidance method using LiDAR | |
| CN115143970A (en) | Obstacle avoidance method and system of underwater vehicle based on threat degree evaluation | |
| CN114967694A (en) | Mobile robot collaborative environment exploration 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 |