Task optimization scheduling method and system for multiple carrying robots

Abstract

The invention provides a multi-carrying robot task optimization scheduling method and system. The method comprises the following steps: establishing a target function; establishing a robot task optimization scheduling model; obtaining a plurality of feasible multi-robot scheduling schemes; a novel multi-fusion cross non-dominated sorting genetic algorithm is adopted to solve an optimal solution; an optimal multi-robot scheduling scheme corresponding to the optimal solution is sent to a corresponding target robot, so that the target robot executes a task according to the optimal multi-robot scheduling scheme; the dispatching system comprises a main control unit, a sensing positioning module and a dispatching robot. The beneficial effects of the invention are that an improved genetic algorithm is provided, a multi-fusion cross strategy, non-dominated sorting and a special crowding degree calculation method are adopted, the diversity and convergence performance of the algorithm are enhanced, and the algorithm is more excellent in stability compared with other algorithms; and a more excellent Pareto optimal solution can be obtained.

Description

Technical Field

[0001] The present invention belongs to the field of robot technology, and in particular relates to a multi-carrier robot task optimization scheduling method and system. Background Art

[0002] In large indoor and outdoor places such as airport terminals, aprons or warehouses, there are problems of optimizing the scheduling of people or objects. Due to the large flow of passengers or goods and their scattered distribution, how to meet the requirements of multiple transport robots to efficiently transport multiple people or goods to the designated location under the conditions of speed, acceleration and range constraints within the activity range of a given transport robot, and to achieve the optimal scheduling of multiple transport robots is a key problem that needs to be solved. At present, there are some common problems in the scheduling schemes of multiple transport robots on the market. For example, it is easy to fall into the problem of local optimal solution, and the convergence speed is usually slow, and the global optimal solution cannot be found quickly; when the non-dominated sorting genetic algorithm is used for non-dominated sorting operations, the neighboring solutions of a solution can be identified from another Pareto optimal solution set, and the congestion accuracy is poor; the crossover and mutation parameters of the algorithm cannot be adaptively adjusted according to the evolution process of the problem, making the algorithm lack of adaptability; only the shortest distance traveled by the robot is used as the objective function, and the shortest working time of the robot and the minimum energy consumption during work are not considered.

[0003] Therefore, in response to the above problems, a method and system for optimizing the scheduling of multiple transport robot tasks is needed, which can meet the requirements of quickly responding to passengers with the shortest distance, the shortest time, and the lowest energy consumption of the transport robots. Summary of the invention

[0004] In order to solve the above technical problems, the present invention provides a method and system for optimizing the scheduling of tasks of multiple transport robots, which is particularly suitable for the optimization scheduling problem of transport robots transporting objects in airport terminals, aprons or warehouses.

[0005] The technical solution adopted by the present invention is: in the first aspect, a method for optimizing the scheduling of tasks of multiple transport robots is provided, comprising:

[0006] A first objective function is constructed based on the distance traveled by each robot in performing the task, a second objective function is constructed based on the time consumed by each robot in performing the task, and finally a third objective function is constructed based on the energy consumed by each robot in performing the task;

[0007] Establish a robot task optimization scheduling model;

[0008] Obtain several feasible multi-robot scheduling solutions;

[0009] A new multi-fusion crossover non-dominated sorting genetic algorithm is used to find the optimal solution;

[0010] The optimal multi-robot scheduling plan corresponding to the optimal solution is sent to the corresponding target robot, so that the target robot performs the task according to the optimal multi-robot scheduling plan.

[0011] Furthermore, the expression of the first objective function is:

[0012]

[0013] Among them, the Manhattan distance Route represents the minimum path length for each transport robot to reach the location of the transported object;

[0014] The second objective function expression is:

[0015]

[0016] The moving speed of the i-th transport robot to carry the j-th transport object is V = (v ji ) m×n to express;

[0017] The third objective function expression is:

[0018]

[0019] Among them, U and I represent the voltage and current of the carrier robot when performing tasks.

[0020] Furthermore, the normalization factor of the first objective function The expression of is,

[0021] Normalization factor of the second objective function The expression of is,

[0022] Normalization factor of the third objective function The expression of is,

[0023] Furthermore, the objective function of the multi-carrier robot task optimization scheduling model is:

[0024]

[0025] Among them, R f V is an environmental factor used to simulate real large indoor or outdoor places. max is the free travel speed of the carrying robot, V = (v ji ) m×n is the current speed of the transport robot, R s is the saturation density of passengers or objects in the environment.

[0026] Furthermore, a novel multi-fusion crossover non-dominated sorting genetic algorithm is used to solve the optimal solution, including the following steps:

[0027] Enter the number of transport robots and transport objects;

[0028] Initialize the population and parameters;

[0029] Calculate the Manhattan distance matrix;

[0030] Calculate the fitness value;

[0031] Perform non-dominated sorting and calculate the special congestion degree based on clustering algorithm;

[0032] Begin iterations and select mating individuals through a tournament;

[0033] A multi-fusion crossover strategy is used for crossover and mutation operations, and a dynamic genetic coefficient strategy is used to dynamically adjust the probability of crossover and mutation;

[0034] Merge the offspring population and the parent population;

[0035] The new population is sorted again by non-dominated sorting, and the special crowding degree based on the clustering algorithm is calculated;

[0036] Conduct elite retention to select outstanding individuals;

[0037] After reaching the maximum number of iterations, the optimal solution is output.

[0038] Furthermore, the special crowding degree based on the clustering algorithm refers to combining the special crowding degree with the clustering algorithm k-means to achieve the goal of only identifying neighboring solutions of the same category of solutions.

[0039] Furthermore, the multi-fusion crossover strategy refers to combining the Levy distribution with the SBX crossover operator and the BLX-a hybrid crossover operator respectively, and performing crossover and mutation operations through the hybrid crossover method.

[0040] In a second aspect, a multi-carrier robot task optimization scheduling system is provided, comprising:

[0041] The main control unit is used to receive the position information of the transport robot and the transported object in an idle state, calculate the optimal path together with the information in the sensing positioning signal, and convert the processed information into a control signal to send to the dispatching robot;

[0042] The sensing and positioning module is used to collect environmental information around the carrier robot, obtain the position information of the carrier robot through algorithm integration of environmental information, and generate sensing and positioning signals to send to the main control unit;

[0043] The dispatching robot is used to display the optimal path on the display of the transport robot in real time according to the received control signal, drive the motor of the transport robot to carry the transport object according to the optimal path to work, and distribute control signals to other transport robots;

[0044] Among them, the scheduling robot performs redundant backup and configures a backup robot to take over the scheduling task.

[0045] Furthermore, the perception and positioning module includes:

[0046] Indoor positioning module, used to collect point cloud data of the surrounding environment of the carrier robot, video images of the surrounding environment, acceleration and posture information in indoor environment;

[0047] The outdoor positioning module is used to collect point cloud data of the surrounding environment of the carrier robot, video images of the surrounding environment, acceleration, posture information and positioning information in an outdoor environment.

[0048] According to a third aspect, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, any of the above methods is implemented.

[0049] The advantages and positive effects of the present invention are as follows: due to the adoption of the above technical scheme, an improved genetic algorithm is proposed, in which a multi-fusion crossover strategy and non-dominated sorting, as well as a special crowding calculation method are adopted, which enhances the diversity and convergence performance of the algorithm, and is also more excellent in stability than other algorithms, and can obtain a better Pareto optimal solution; a multi-fusion crossover scheme integrating SBX crossover, BLX-a crossover, dynamic genetic coefficient and Levy distribution is proposed to avoid falling into a local optimal solution, improve population diversity, and enhance the ability to search the target space and decision space; the Levy distribution is combined with the crossover operator, which can effectively search for local potential individuals without destroying the guidance information, and is helpful to avoid local optimality and improve the diversity of the population; by configuring a backup robot, the stable operation of the system is ensured, and the reliability and safety of the scheduling task are improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 is a flow chart of a scheduling method according to an embodiment of the present invention

[0051] Figure 2 This is a flow chart of the NMFCNSGA-II algorithm according to an embodiment of the present invention.

[0052] Figure 3 is an algorithm parameter setting table of an embodiment of the present invention

[0053] Figure 4This is a comparison result table of the total objective function values ​​of an embodiment of the present invention.

[0054] Figure 5 Schematic diagram of the scheduling system structure of an embodiment of the present invention DETAILED DESCRIPTION

[0055] In the following description, those skilled in the art will recognize that the embodiments of the invention described below can be implemented in various ways (e.g., processes, devices, systems, equipment, or methods) on a tangible computer-readable medium. Throughout this discussion, components may be described as separate functional units (which may include subunits), but those skilled in the art will recognize that various components or parts thereof may be divided into separate components, or may be integrated together (including integration into a single system or component). It should be noted that the functions or operations discussed herein may be implemented as components. Components may be implemented in software, hardware, or a combination thereof.

[0056] The embodiments of the present invention are described below with reference to the accompanying drawings.

[0057] like Figure 1 As shown, the present invention provides a multi-carrier robot task optimization scheduling method, comprising:

[0058] S101, constructing a first objective function based on the distance traveled by each robot in performing a task, constructing a second objective function based on the time consumed by each robot in performing a task, and finally constructing a third objective function based on the energy consumed by each robot in performing a task;

[0059] S102, establishing a robot task optimization scheduling model;

[0060] S103, obtaining several feasible multi-robot scheduling solutions;

[0061] S104, using a novel multi-fusion crossover non-dominated sorting genetic algorithm to find the optimal solution;

[0062] An improved genetic algorithm NMFCNSGA-II is proposed, which adopts multi-fusion crossover strategy and non-dominated sorting, as well as a special crowding calculation method, which enhances the diversity and convergence performance of the algorithm. It is also better in stability than other algorithms and can obtain better Pareto optimal solutions.

[0063] S105. Send the optimal multi-robot scheduling plan corresponding to the optimal solution to the corresponding target robot, so that the target robot performs the task according to the optimal multi-robot scheduling plan.

[0064] In some embodiments, the first objective function expression is:

[0065]

[0066] Among them, the Manhattan distance Route represents the minimum path length for each transport robot to reach the location of the transported object;

[0067] Assume that the total number of transport robots is N, represented by a set R = {R1, R2, …, R n}, the total number of transport robots that can be dispatched at the current moment is n; the total number of transport objects is M, using the set P = {P1, P2, ..., P m} indicates that the total number of objects being transported at the current moment is m. At the same moment, a transport robot can only carry one unit of transport object. The current state of the robot is represented by the set S = {S1, S2, …, S i} indicates that S i =1 means robot i is currently carrying an object, S i = 0 means robot i is in idle state. The current position of object j is P j =(x j ,y j ), the current position of the transport robot i is R j =(x j ,y j ). Under the premise of avoiding collision between the transport robot and obstacles, a path network set of people or goods and the transport robot is established Web = {Web1, Web2, …, Web n},in The minimum path length for each carrying robot to reach the location of the carrying object is the Manhattan distance Route, which is expressed by the total number of grids on the path multiplied by the length cost E of the robot moving one grid, and the expression is:

[0068] Route=min[n(Web1)∪n(Web2)…∪n(Web p )]E

[0069]

[0070] in, They represent the grid positions of the i-th transport robot and the p-th transport robot at time T respectively; Respectively represent the grid positions of the i-th carrier robot and the p-th carrier robot at time T+1; n(Web p ) represents the number of grids on the path of the p-th transport robot. The Manhattan distance matrix between each transport robot and each passenger or cargo is expressed by:

[0071]

[0072] The second objective function expression is:

[0073]

[0074] The moving speed of the i-th transport robot to carry the j-th transport object is V = (v ji ) m×n to express;

[0075] Set the initial speed v0 and final speed v of the carrying robot r All are 0m / s, the maximum speed v max According to the Manhattan distance matrix R between each carrying robot and each carrying object, the moving speed V of each carrying robot carrying each carrying object is calculated in turn. ji represents the speed of the i-th transport robot traveling to the j-th transport object, and the time matrix between each transport robot and each transport object is expressed by T:

[0076]

[0077] The third objective function expression is:

[0078]

[0079] Among them, U and I represent the voltage and current of the carrier robot when performing tasks.

[0080] Energy consumption of the transport robot E r Divided into the energy consumption of the equipment itself E d and the mobile energy consumption E of the robot during operation m Two parts. The moving energy consumption E of the carrying robot during operation m =UI×t. Considering the continuous operation of the transport robot in large indoor or outdoor places, the energy consumption power of the transport robot’s own equipment is defined as P d , then the energy consumption of the robot device itself is E d =P d × t, we get the total energy consumption E of the transport robot. r =E d +E m .

[0081] In some embodiments, the normalization factor of the first objective function is The expression of is,

[0082] Normalization factor of the second objective function The expression of is,

[0083] Normalization factor of the third objective function The expression of is,

[0084] In some embodiments, the objective function of the multi-carrier robot task optimization scheduling model is:

[0085]

[0086] Among them, R f V is an environmental factor used to simulate real large indoor or outdoor places. max is the free travel speed of the carrying robot, V = (v ji ) m×n is the current speed of the transport robot, R s is the saturation density of passengers or objects in the environment.

[0087] F1, F2, and F3 represent the total moving distance of the transport robot, the total time of carrying passengers or goods, and the total energy consumption during operation. ji ) m×n A variable that is 0 or 1, g ji = 0 means that the jth transport object is not carried by the i-th transport robot, g ji =1 means that the jth transport object is carried by the i-th transport robot; define [Length, Wide] as the moving range of the transport robot, Length and Wide represent the length and width of the moving range respectively. The constraints of the objective function are as follows:

[0088]

[0089] In some embodiments, Figure 2 As shown, a new multi-fusion crossover non-dominated sorting genetic algorithm is used to solve the optimal solution, including the following steps:

[0090] S201, input the number of transport robots and transport objects;

[0091] S202, initializing population and parameters;

[0092] S203, calculating the Manhattan distance matrix;

[0093] S204, calculating fitness value;

[0094] S205, performing non-dominated sorting and calculating a special congestion degree based on a clustering algorithm;

[0095] The non-dominated sorting method based on the special crowding degree of the clustering algorithm is divided into four steps. The first step is to initialize the dominance level of all individuals in the population to 0; the second step is to use the fast non-dominated sorting method to sort all individuals, and calculate the crowding values ​​of each layer of individuals in the decision space domain and the object space domain respectively; the third step is to calculate the special crowding distance based on the clustering algorithm according to the crowding values ​​in the decision space domain and the object space domain; the fourth step is to sort the individuals of the same level according to the special crowding distance based on the clustering algorithm, and select the individuals with higher crowding as the next mating individuals. The crowding degree is calculated based on the result of sorting the non-dominated solution. The crowding degree indicates the density of the surrounding individuals. The higher the crowding degree, the more crowded the surrounding individuals are. Therefore, the individuals with higher crowding degree are more likely to be selected at the same level. First, sort the individuals at the dth level in order from the maximum value to the minimum value. The crowding degree of the first and last individuals is set to 1, and the crowding degree of other individuals is based on Calculated.

[0096] S206, start iteration, and select mating individuals through the tournament;

[0097] Randomly select two individuals from the population for comparison, and select the better individual as part of the mating pool. Repeat this process until a sufficient number of individuals are generated to enter the mating pool.

[0098] S207, using a multi-fusion crossover strategy to perform crossover and mutation operations, and dynamically adjusting the probability of crossover and mutation through a dynamic genetic coefficient strategy;

[0099] S208, merging the offspring population and the parent population;

[0100] S209, performing non-dominated sorting on the new population again, and calculating the special congestion degree based on the clustering algorithm;

[0101] S210, conduct elite retention and select outstanding individuals;

[0102] S211. Output the optimal solution after reaching the maximum number of iterations.

[0103] In order to verify the effectiveness and superiority of the NMFCNSGA-II algorithm, the NMFCNSGA-II algorithm was compared with three algorithms: NNSGA-II (SBX), NNSGA-II (BLX) and NSGA-II for solving the multi-carrier robot optimization scheduling model. The parameter settings of all algorithms are as follows Figure 3 The comparison results of the multi-carrier robot optimization scheduling model in terms of the average value, standard deviation, optimal value, and worst value of the total objective function value are shown in Figure 4As shown. The smaller the total objective function value, the better. The bold data represents the best result of the algorithm solving the total objective function value of the optimization scheduling model. From the comparison results of NSGA-II and NNSGA-II (SBX), it can be seen that the total objective function value becomes smaller, which shows that the non-dominated sorting based on the special crowding distance of the clustering algorithm, the dynamic genetic coefficient strategy and the Levy distribution fusion strategy can effectively improve the performance of the algorithm, enhance the adaptability of the algorithm to avoid falling into the local optimal situation, and obtain a better Pareto optimal solution. In terms of stability, NNSGA-II (SBX) has also improved compared with NSGA-II, but the effect is not obvious. In view of the problem that the stability of NNSGA-II (SBX) is not significantly improved. By comparing the data results of NNSGA-II (SBX) and NNSGA-II (BLX) under different crossover strategies. It can be seen that when the dynamic genetic coefficient strategy and the Levy distribution fusion strategy are used at the same time and the non-dominated sorting based on the special crowding distance of the clustering algorithm, the total objective function value of the BLX-a crossover strategy is not much different from the SBX crossover strategy, but the stability of the algorithm is significantly improved.

[0104] In some embodiments, the special crowding degree based on the clustering algorithm refers to combining the special crowding degree with the clustering algorithm k-means to achieve the goal of only identifying neighboring solutions of the same category of solutions.

[0105] Special crowding distance (SCD) is a distance metric improved on the basis of crowding distance (CD), which can reflect the crowding degree of individuals and is widely used as an indicator to judge the ranking size of individuals with the same non-dominated degree in the non-dominated sorting operation of most multimodal multi-objective algorithms. However, in the decision space, the area between a solution and its neighboring solutions cannot be close to the true PS, or the neighboring solutions of one solution can be identified from the PS of another. In order to solve the shortcomings of the above-mentioned SCD method, the clustering algorithm k-means can be combined with special crowding to form the CSCD method. In the CSCD method, the k-means algorithm is able to divide solutions with the same non-dominated level into multiple categories. In addition, the individuals in each category also belong to the same area, and the CSCD method only identifies the neighbors of the same category of solutions, which improves the accuracy of the crowding distance.

[0106] In some embodiments, the multi-fusion crossover strategy refers to combining the Levy distribution with the SBX crossover operator and the BLX-a hybrid crossover operator, and performing crossover and mutation operations in a hybrid crossover manner.

[0107] The expression of Levy distribution is In order to avoid premature convergence and improve population diversity, the expression of the Levy distribution fusion scheme that integrates the crossover operator and the Levy distribution is described as follows: Among them, F1 and F2 are the two selected parents, and μ represents the crossover probability coefficient. In the two crossover algorithms, the SBX crossover operator is a crossover operator that simulates a single-point binary crossover. Assume that the two r-dimensional parent individuals are as follows: and and is a new individual generated by crossover, according to the formula The next generation can be obtained by calculation. BLX-a is a linear hybrid crossover operator, and its crossover method formula is Preferably, the hybrid crossover modes include but are not limited to dimension-based hybrid crossover, solution-based hybrid crossover, dynamic adaptive hybrid crossover and random selection crossover.

[0108] A multi-fusion crossover scheme integrating SBX crossover, BLX-a crossover, dynamic genetic coefficient and Levy distribution is proposed to avoid falling into local optimal solutions, improve population diversity, and enhance the ability to search target space and decision space. Combining Levy distribution with crossover operator can effectively search local potential individuals without destroying the guidance information, and help avoid local optimal solutions and improve population diversity.

[0109] An embodiment of the present application also provides a multi-carrier robot task optimization scheduling system.

[0110] like Figure 3 As shown, a multi-carrier robot task optimization scheduling system includes:

[0111] The main control unit 300 is used to receive the position information of the transport robot and the transport object in an idle state, calculate the optimal path together with the information in the sensing positioning signal, and convert the processed information into a control signal to be sent to the dispatching robot 500;

[0112] Preferably, the position information of the carrying robot and the carrying object in an idle state is sent by the server to the main control unit.

[0113] The sensing and positioning module 400 is used to collect environmental information around the carrier robot, obtain the position information of the carrier robot by fusing the environmental information through an algorithm, and generate a sensing and positioning signal to send to the main control unit 300;

[0114] Preferably, the Kalman or extended Kalman filter algorithm is used to fuse the environmental information to obtain the position information of the carrier robot.

[0115] The dispatching robot 500 is used to display the optimal path in real time on the display of the transporting robot according to the received control signal, drive the motor of the transporting robot to carry the object according to the optimal path, and distribute control signals to other transporting robots. The dispatching robot 500 is redundantly backed up, and a backup robot is configured to take over the dispatching task.

[0116] By configuring backup robots, the stable operation of the system is ensured and the reliability and safety of scheduling tasks are improved.

[0117] In some embodiments, the sensing and positioning module includes:

[0118] Indoor positioning module 401, indoor positioning module, used to collect point cloud data of the surrounding environment of the carrier robot, video images of the surrounding environment, acceleration and posture information in an indoor environment;

[0119] Preferably, the acceleration and posture information of the carrying robot is collected by IMU, the 2D or 3D point cloud data of the surrounding environment of the carrying robot is collected by radar, and the video image of the surrounding environment is collected by visual camera.

[0120] IMU can provide data such as acceleration and attitude, which can be used to improve the robustness and accuracy of positioning. Radar is used to collect 2D or 3D point cloud data of the surrounding environment of the carrier robot to provide data for environmental information perception. The core principle of laser SLAM (Simultaneous Localization and Mapping) is to use laser sensors to receive surrounding environmental information and fuse this information through algorithms to achieve self-positioning and map construction of the robot.

[0121] The outdoor positioning module 402 is used to collect point cloud data of the surrounding environment of the carrier robot, video images of the surrounding environment, acceleration, posture information and positioning information in an outdoor environment.

[0122] Preferably, the acceleration and posture information of the carrying robot is collected through IMU, the positioning information of the carrying robot is collected through GNSS sensor and base station positioning, the 2D or 3D point cloud data of the surrounding environment of the carrying robot is collected through radar, and the video image of the surrounding environment is collected through visual camera.

[0123] IMU can provide data such as acceleration and attitude, which can be used to improve the robustness and accuracy of positioning. Radar is used to collect 2D or 3D point cloud data of the surrounding environment of the carrier robot to provide data for environmental information perception. The core principle of laser SLAM (Simultaneous Localization and Mapping) is to use laser sensors to receive surrounding environmental information and fuse this information through algorithms to achieve self-positioning and map construction of the robot. Sensors such as Beidou / GPS receive data sent by satellites through wireless communication modules and provide positioning information for the carrier robot through base station positioning.

[0124] An embodiment of the present application further provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, any of the above methods is implemented.

[0125] The embodiments of the present invention are described in detail above, but the contents described are only preferred embodiments of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. A multi-carrier robot task optimization scheduling method, characterized in that: include: A first objective function is constructed based on the distance traveled by each robot in performing the task, a second objective function is constructed based on the time consumed by each robot in performing the task, and finally a third objective function is constructed based on the energy consumed by each robot in performing the task; Establish a robot task optimization scheduling model; Obtain several feasible multi-robot scheduling solutions; A new multi-fusion crossover non-dominated sorting genetic algorithm is used to find the optimal solution; The optimal multi-robot scheduling plan corresponding to the optimal solution is sent to the corresponding target robot, so that the target robot performs the task according to the optimal multi-robot scheduling plan.

2. The multi-carrier robot task optimization scheduling method according to claim 1 is characterized in that: The expression of the first objective function is: Among them, the Manhattan distance Route represents the minimum path length for each transport robot to reach the location of the transported object; The second objective function expression is: The moving speed of the i-th transport robot to carry the j-th transport object is V = (v ji ) m×n to express; The third objective function expression is: Among them, U and I represent the voltage and current of the carrier robot when performing tasks.

3. The multi-carrier robot task optimization scheduling method according to claim 2 is characterized in that: Normalization factor of the first objective function The expression of is, Normalization factor of the second objective function The expression of is, Normalization factor of the third objective function The expression of is, 4. The multi-carrier robot task optimization scheduling method according to claim 1 is characterized in that: The objective function of the multi-carrier robot task optimization scheduling model is: Among them, R f V is an environmental factor used to simulate real large indoor or outdoor places. max is the free travel speed of the carrying robot, V = (v ji ) m×n is the current speed of the transport robot, R s is the saturation density of passengers or objects in the environment.

5. The multi-carrier robot task optimization scheduling method according to claim 1 is characterized in that: A new multi-fusion crossover non-dominated sorting genetic algorithm is used to solve the optimal solution, including the following steps: Enter the number of transport robots and transport objects; Initialize the population and parameters; Calculate the Manhattan distance matrix; Calculate the fitness value; Perform non-dominated sorting and calculate the special congestion degree based on clustering algorithm; Begin iterations and select mating individuals through a tournament; A multi-fusion crossover strategy is used for crossover and mutation operations, and a dynamic genetic coefficient strategy is used to dynamically adjust the probability of crossover and mutation; Merge the offspring population and the parent population; The new population is sorted again by non-dominated sorting, and the special crowding degree based on the clustering algorithm is calculated; Conduct elite retention to select outstanding individuals; After reaching the maximum number of iterations, the optimal solution is output.

6. The multi-carrier robot task optimization scheduling method according to claim 5 is characterized in that: The special crowding degree based on clustering algorithm refers to combining the special crowding degree with the clustering algorithm k-means to only identify neighboring solutions of the same category.

7. The multi-carrier robot task optimization scheduling method according to claim 5 is characterized in that: The multi-fusion crossover strategy refers to combining the Levy distribution with the SBX crossover operator and the BLX-a hybrid crossover operator, and performing crossover and mutation operations through hybrid crossover.

8. A multi-carrier robot task optimization scheduling system, characterized in that: include: The main control unit is used to receive the position information of the transport robot and the transported object in an idle state, calculate the optimal path together with the information in the sensing positioning signal, and convert the processed information into a control signal to send to the dispatching robot; The sensing and positioning module is used to collect environmental information around the carrier robot, obtain the position information of the carrier robot through algorithm integration of environmental information, and generate sensing and positioning signals to send to the main control unit; The dispatching robot is used to display the optimal path on the display of the transport robot in real time according to the received control signal, drive the motor of the transport robot to carry the transport object according to the optimal path to work, and distribute control signals to other transport robots; Among them, the scheduling robot performs redundant backup and configures a backup robot to take over the scheduling task.

9. The multi-carrier robot task optimization scheduling system according to claim 7, characterized in that: The perception and positioning module includes: Indoor positioning module, used to collect point cloud data of the surrounding environment of the carrier robot, video images of the surrounding environment, acceleration and posture information in indoor environment; The outdoor positioning module is used to collect point cloud data of the surrounding environment of the carrier robot, video images of the surrounding environment, acceleration, posture information and positioning information in an outdoor environment.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images