Modular robotic reconfiguration planning method and system

Abstract

The application provides a modular robot reconfiguration planning method and system, and relates to the technical field of robot control, and comprises the following steps: aiming at three optimization targets of the least number of mobile modules, the least number of disconnected connecting mechanisms and the shortest moving distance, global optimal solution sets of each optimization target are solved; the global optimal solution sets of each optimization target are respectively divided into hierarchical sequence habitats, and a plurality of microhabitats are generated for each optimization target; for each microhabitat, common features of the reconfiguration solution are extracted and coded into common chromosome fragments; based on the common chromosome fragments, non-dominated sorting iterative optimization is carried out in the habitat through a genetic algorithm until the maximum number of iterations is reached, the final population of all habitats is combined, and the non-dominated solution is extracted as the optimal solution output. Through a hierarchical sequence habitat division-based genetic algorithm, the application solves the solving of the three coupled optimization targets in the reconfiguration process and searches for the Pareto front of the optimal reconfiguration problem.

Description

Technical Field

[0001] This invention relates to the field of robot control technology, and specifically to a modular robot configuration planning method and system. Background Technology

[0002] A mobile modular robot is an assembly of multiple mobile modules connected by docking mechanisms. Each module possesses independent sensing, control, and mobility capabilities. The assembly changes its configuration through a metamorphic process to adapt to diverse environments and tasks. Metamorphism is one of the core capabilities of modular robots, enabling them to switch configurations to suit different terrains or transport various objects. Efficient and low-cost metamorphism ensures the robot's flexibility and stability in dynamic environments and tasks, and enhances its endurance.

[0003] Modular robot reconfiguration planning technology still faces challenges: Existing reconfiguration planning studies often use movement distance and reconfiguration steps as indicators of reconfiguration cost, but reconfiguration cost encompasses more dimensions; the number of moving modules directly affects movement cost; for active docking and locking devices, fewer actions result in lower energy consumption and a lower error rate, but existing research rarely considers these two factors; in addition, multi-indicator optimization has coupling relationships, improving one indicator often leads to the deterioration of another, and the optimal solution exists in the form of a Pareto front; the coupling of multiple optimization objectives increases the complexity of reconfiguration planning, and the number of reconfiguration schemes increases exponentially with the number of modules, and the solution space expands factorially with the number of modules; searching for Pareto optimal solutions in a large-scale solution space faces enormous challenges.

[0004] In summary, existing modular robot configuration planning technologies suffer from problems such as strong multi-objective coupling, solution space explosion, and incomplete cost indicators. Summary of the Invention

[0005] To address the aforementioned problems, this invention proposes a modular robot reconfiguration planning method and system. It employs a genetic algorithm based on hierarchical sequence habitat partitioning to solve the three coupled optimization objectives during the reconfiguration process and search for the Pareto front of the optimal reconfiguration problem.

[0006] According to some embodiments, the present invention adopts the following technical solution:

[0007] A modular robot variant planning method includes:

[0008] For the three optimization objectives of minimizing the number of mobile modules, minimizing the number of disconnected connection mechanisms, and minimizing the travel distance, the global optimal solution set for each optimization objective is solved.

[0009] The global optimal solution set for each optimization objective is divided into hierarchical sequential habitats, generating several sub-habitats for each optimization objective, with each sub-habitat containing several variant solutions;

[0010] For each niche, common features of its allotropic solutions are extracted and encoded as common chromosome segments;

[0011] Based on a common chromosome segment, a genetic algorithm is used to perform non-dominated sorting iterative optimization within the habitat until the maximum number of iterations is reached. The final populations of all habitats are then merged, and the non-dominated solution is extracted as the optimal solution for output.

[0012] According to some embodiments, the present invention adopts the following technical solution:

[0013] A modular robot configuration planning system, comprising:

[0014] The separate solution modules are configured to solve for the global optimal solution set for each of the three optimization objectives: minimizing the number of moving modules, minimizing the number of disconnected connecting mechanisms, and minimizing the moving distance.

[0015] The habitat partitioning module is configured to perform hierarchical sequential habitat partitioning on the global optimal solution set of each optimization objective, generating several sub-habitats for each optimization objective, with each sub-habitat containing several variant solutions;

[0016] The feature encoding module is configured to extract common features of the allotropic solutions for each niche and encode them as common chromosome segments.

[0017] The genetic optimization module is configured to: based on common chromosome segments, perform non-dominated sorting iterative optimization within the habitat using a genetic algorithm until the maximum number of iterations is reached, merge the final populations of all habitats, and extract the non-dominated solutions as the optimal solution output.

[0018] According to some embodiments, the present invention adopts the following technical solution:

[0019] A computer program product includes a computer program that, when executed by a processor, implements the modular robot variant planning method.

[0020] According to some embodiments, the present invention adopts the following technical solution:

[0021] A non-transitory computer-readable storage medium is provided for storing computer instructions, which, when executed by a processor, implement the modular robot variant planning method described above.

[0022] According to some embodiments, the present invention adopts the following technical solution:

[0023] An electronic device includes a processor, a memory, and a computer program; wherein the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory to enable the electronic device to implement the modular robot variant planning method.

[0024] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0025] This invention introduces hierarchical sequence habitat partitioning into the metamorphic optimization programming problem. It constructs habitat partitioning based on the global optimal solution set of each optimization objective, and performs iterative optimization within the habitat using a genetic algorithm. This effectively focuses the search direction near the Pareto front, significantly reducing the search space and overcoming the problem of getting trapped in local optima.

[0026] The niche partitioning method of the present invention can ensure that the Pareto optimal solution set is located in each niche. Through optimization within the niche, it can approximate all Pareto solutions without missing any optimal solutions, and has the comprehensiveness to cover all optimal solutions.

[0027] The technology of this invention has scalability; the hierarchical sequence habitat partitioning method can be extended to other multi-objective optimization problems, improve Pareto front approximation ability, and achieve higher quality, multi-objective equal optimization search capabilities. Attached Figure Description

[0028] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0029] Figure 1 This is a flowchart of the method in Example 1.

[0030] Figure 2 This is a flowchart of the calculation of common chromosome segments in Example 1.

[0031] Figure 3 This is a rectangular area diagram between the mobile module location and the corresponding idle location in Example 1.

[0032] Figure 4 This is a diagram of the component identification process in Example 1.

[0033] Figure 5 This is a diagram illustrating the calculation process of the index value in Example 1. Detailed Implementation

[0034] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0035] It should be noted that the following detailed descriptions are exemplary and intended to provide further illustration of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0036] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of exemplary embodiments according to the invention. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0037] Example 1

[0038] One embodiment of the present invention provides a modular robot variant planning method, comprising:

[0039] Step 1: For the three optimization objectives of minimizing the number of moving modules, minimizing the number of disconnected connecting mechanisms, and minimizing the moving distance, solve for the global optimal solution set for each optimization objective;

[0040] Step 2: Perform hierarchical sequence habitat partitioning on the global optimal solution set of each optimization objective, generating several sub-habitats for each optimization objective, with each sub-habitat containing several variant solutions;

[0041] Step 3: For each niche, extract the common features of its allotropic solutions and encode them as common chromosome segments;

[0042] Step 4: Based on the common chromosome segment, perform non-dominated sorting iterative optimization within the habitat using a genetic algorithm until the maximum number of iterations is reached. Then, merge the final populations of all habitats and extract the non-dominated solution as the optimal solution output.

[0043] As an example, based on the non-dominated sorting genetic algorithm NSGA-III, a genetic algorithm based on hierarchical sequence habitat partitioning is proposed to search for the Pareto front of the optimal metamorphic problem. This algorithm solves the problem of solving three coupled optimization objectives in the metamorphic process. The three objectives are to minimize the number of moving modules, minimize the number of disconnected connecting mechanisms, and minimize the moving distance.

[0044] In this embodiment, the modular robot must meet the following conditions:

[0045] (1) The robot consists of multiple modules, all of which are isomorphic. Two combinations of modules in the same position are considered to be the same configuration.

[0046] (2) The module has a regular shape and a definite connection direction, such as a rectangular module with four-sided connection and a hexagonal module with six-way connection. The method of this embodiment is not applicable to irregular modules and robot platforms without a fixed connection direction, such as irregular modules and spherical modules with arbitrary direction connection.

[0047] In this method, each individual in the population represents a feasible solution to the problem, i.e., a feasible allosteric solution or allosteric scheme. Individuals are represented by chromosome encoding: for a population containing... The modular transformation process assigns 1 to each point in the target configuration. The target ID, in the variant scheme, is equivalent to assigning a corresponding target ID to the module, i.e., the target point; let... For inclusion The first individual in the population chromosome number 1, i.e., the 1st chromosome One feasible solution. yes 3D row vectors It consists of two parts:

[0048] Part One contains numbers 1 through 1. Each element, namely This section is used to indicate the target assignment of a module. The column number in this section represents the target ID, and the element value represents the module ID, meaning that the module will be moved to the position corresponding to the target ID.

[0049] The other part is the alignment point, which is determined by the first... element This means that the alignment point determines the relative position of the initial configuration and the target configuration in the same coordinate system, and is used for path and distance calculations.

[0050] In this method, a niche is a group of individuals sharing common characteristics; that is, a niche consists of allotropic solutions with common characteristics. Within a niche, common characteristics refer to the similarity among individuals determined by shared chromosome segments and reflected in the three optimization objective indicators. For example, individuals belonging to the niche with the fewest movement modules all represent allotropic solutions with the same few movement modules, while differing in the other two optimization objectives. In the solution space, individuals within the same niche cluster together in a region, which is the niche. The entire population is divided into multiple small regions through multiple niches, and each individual belongs to a unique niche.

[0051] Under the above conditions and definitions, a modular robot variant planning method, such as... Figure 1 As shown, the steps are as follows:

[0052] Step 1: For the three optimization objectives of minimizing the number of mobile modules, minimizing the number of disconnected connection mechanisms, and minimizing the travel distance, solve for the global optimal solution set for each optimization objective.

[0053] For the optimization objective of minimizing the number of moving modules, the target configuration is placed above the initial configuration and slid row by row and column by column. The overlapping parts of the two configurations are checked, and the sliding step with the most overlapping modules is found. The overlapping parts in this step are regarded as stationary parts and do not move during the configuration change process; the remaining modules are moving parts, and the number of modules contained in them is the minimum number of moving modules.

[0054] For the optimization objective of minimizing the number of disconnections in the connecting mechanism, the search is performed using an iterative matching and splitting algorithm based on heuristic search and a splitting tree structure, specifically:

[0055] In the tree structure, nodes represent the derived variant components. The initial configuration is set as the root node of the tree. The connection between parent and child nodes represents the next step of splitting the remaining part based on the child node components after the parent node component has been split in one step. The component with the minimum heuristic cost is searched as the next node and added to the search tree until the initial configuration is completely split, thus completing the search. Each branch of the tree represents a splitting scheme, i.e., a variant solution. The branch with the fewest disconnected connections constitutes the global optimal solution set. Among them, the variant component (or component) is a subconfiguration that exists simultaneously in the initial configuration and the target configuration, and the module number and arrangement shape are completely consistent. During the variant process, the initial configuration is split into multiple components. After these components move to the target position, they are reassembled into the target configuration. Each component contains at least one module.

[0056] For the optimization objective of minimizing the movement distance, a near-optimal estimation method is adopted. Specifically, the target configuration is placed above the initial configuration and slid row by row and column by column. The overlapping parts of the two configurations are checked. During each sliding step, the minimum total movement distance from multiple modules to multiple target positions is calculated using a greedy algorithm. The sliding step with the minimum total movement distance is found among all sliding steps and is taken as the global near-optimal solution set with the shortest movement distance.

[0057] Step 2: Divide habitats based on the global optimal solution set.

[0058] Habitats include four types: the globally optimal solution set corresponding to the minimum number of moving modules. Habitat class; the globally optimal solution set corresponding to the minimum number of disconnected connection mechanisms. Habitat category; the globally optimal solution set corresponding to the shortest movement distance Habitat class; Non-optimal solution set correspondence Habitat category.

[0059] in, , and Habitat categories all contain multiple subhabitats, and The habitat category contains only one small habitat.

[0060] Step 3: Identify the common chromosome segments among the populations within each habitat class, such as... Figure 2 As shown, specifically:

[0061] (1) For the global optimal solution set with the fewest moving modules, extract the distribution location and module number of the stationary modules in each solution. Solutions with the same distribution location and stationary module number are grouped together to form a small habitat; different groups form Multiple subhabitats under the habitat category.

[0062] In each niche, the target ID of the quiescent module in the target configuration has been determined. The quiescent module ID is then filled into the chromosome. In the dimensional row vector, the corresponding target ID serves as the column number of the chromosome row vector, and the module ID is filled into the corresponding column as a value, forming a fixed common chromosome segment in the chromosome. The common chromosome segments of different habitats are different, which is the only difference between different habitats.

[0063] (2) For the global optimal solution set with the fewest disconnections in the connecting mechanism, use Euclidean distance to calculate the two solutions. and solution The gap between The calculation formula is as follows:

[0064]

[0065] After calculating the distance, based on the principle that solutions with similar distances belong to the same cluster, multiple groups are formed using the K-means clustering algorithm.

[0066] Specifically, firstly, k solutions are randomly selected from the global optimal solution set as initial cluster centers. The distance from all solutions in the global optimal solution set to each cluster center is calculated, and each solution is assigned to the nearest cluster center to form k initial clusters. Then, the mean of all samples in each cluster is recalculated to update the center of the cluster. The assignment and update steps are repeated until the position of the cluster centers no longer changes significantly, and the clustering result, i.e., grouping, is obtained.

[0067] Each group corresponds to a small habitat. Within each group, the common parts of each solution are extracted by chromosome comparison and used as common chromosome segments.

[0068] (3) For the global optimal solution set with the shortest movement distance, the solutions in the solution set with the same alignment point form a habitat.

[0069] By using the alignment point element values ​​in the solution, the initial configuration and the target configuration are aligned and overlapped in the same matrix, thus determining the mobile module in the initial configuration and its corresponding free position in the target configuration;

[0070] Each mobile module location and its corresponding available location form a rectangular area, such as... Figure 3 As shown, the blue blocks represent the overlapping parts of the initial and target configurations, the dashed boxes indicate the target positions for module transformations, and modules of other colors need to be moved to the positions of the dashed boxes of corresponding colors.

[0071] Modules located outside all rectangular regions do not move during the transformation process, such as... Figure 3 As shown in the blue block with a yellow triangle, fill the module IDs, along with the alignment points, that are outside all rectangular areas into the chromosome to form a common chromosome segment.

[0072] (4) For non-optimal solution sets corresponding to For habitats, there are no common chromosome segments among the chromosomes, and the chromosomes within the habitat are set to be completely empty.

[0073] Step 4: Based on the common chromosome segments of each microhabitat, generate the initial population of individuals for each microhabitat.

[0074] First, calculate the population size within the niche. Segments of chromosomes other than the common chromosome segment are variable segments. The number of columns contained in a variable segment determines the size of the solution space to be searched within the niche. For segments containing a certain number of columns... The variable segment of habitat, its population size To adapt to different space sizes, its design is as follows:

[0075]

[0076] in, The population size is generated by the variable segment (excluding the alignment point), and its power exponent parameter is: , This is the population size generated by the variable alignment point; if the alignment point is fixed, it takes the value 0. For scale parameters, This represents the total number of modules.

[0077] Within each small population area, the variable segment filling module ID of the chromosome is used for population initialization, specifically as follows:

[0078] When generating initial individuals in a niche, the common chromosome segment remains unchanged as a fixed segment, while the module IDs of the non-common chromosome segments are filled into the variable segment in a random order.

[0079] Step 5: Perform chromosome crossover and variation calculations within the habitat.

[0080] Targeting domestic Chromosomes, selected Interleaving individuals and selecting Each individual undergoes random mutation to form a new individual. and These are the proportional coefficients for selecting the quantity. Two coefficients are predefined, and their values ​​can be determined through algorithm debugging.

[0081] Step 6: Independently calculate the index values ​​of all individuals on the three optimization objectives, use them as the ternary cost of the modified solution, perform non-dominated sorting on all individuals, and select a new population for the next iteration.

[0082] The index values ​​for the three optimization objectives are calculated using chromosomes, specifically:

[0083] 1. Identify allosteric components from chromosomes. Allosteric components are subconfigurations that exist simultaneously in both the initial and target configurations, and whose module numbering and arrangement are completely identical. Components are identified through a sliding test, such as... Figure 4 As shown, the process of identifying components is as follows:

[0084] Based on the chromosome records, fill the empty slots in the target configuration corresponding to the column number with the module ID; set a OK An empty matrix of columns, where and It is the height and width of the initial configuration. and This refers to the height and width of the target configuration; the initial configuration is placed at the center of the blank matrix to form the sliding base matrix. The alignment point values ​​in the chromosome are then... Convert to row and column coordinates The calculation method is as follows:

[0085]

[0086] in, It is the remainder sign, that is right Take the remainder, with Figure 4 Taking the chromosome as an example, the alignment point in the chromosome is 14, and the row and column coordinates are calculated to be (2, 4) according to the formula.

[0087] Place the target configuration into the sliding base matrix and slide the target configuration step by step in the order shown by the red dashed lines; during the sliding, the upper left corner of the target configuration (e.g., Figure 4 The rectangle with the red pin in the middle slides step by step in the order shown by the red dotted lines. In each sliding step, there are overlapping modules between the target configuration and the initial configuration, such as... Figure 4 As shown in the two-color rectangle; if the IDs of overlapping modules are the same, then the overlapping modules form a variant unit; Figure 4 In the process, when the target configuration slides to position (2, 4) (row 2, column 4), which is the row and column position represented by alignment point 14, the overlapping sub-configuration containing modules 1, 2, and 3 is identified as a single element. The overlapping modules "4|5" and "5|6" below the positions of modules 1, 2, and 3 do not belong to this element because their module IDs are different. Other elements are calculated in other sliding steps, such as... Figure 4 As shown, the other two components are identified at positions (2, 3) (2nd row, 3rd column) and (3, 3) (3rd row, 3rd column), respectively.

[0088] 2. After identifying all components, a variant splitting scheme is obtained; based on the splitting scheme, the index values ​​of the three optimization objectives are calculated, such as... Figure 5 As shown, specifically:

[0089] ① Calculate the number of disconnected connections: Connections within a component are not disconnected; all other connections are counted as disconnected connections. Let... The number of connections in the initial configuration. Chromosomes The Middle The number of connections for each element, then the number of disconnections. The calculation formula is

[0090] (33)

[0091] Where n is the chromosome The number of components in the text.

[0092] ② Calculate the number of moving modules: Since the element containing modules 1, 2, and 3 is identified in the sliding step represented by the alignment point, this element is considered a static element. No movement occurs during the transformation process; all components except the static component need to be moved; if no component is found in the sliding step represented by the alignment point, then all modules need to be moved, and the number of moved modules is... The calculation is as follows:

[0093] (34)

[0094] in, It is a static component A collection of modules, It represents the number of modules contained in a static component. It represents the total number of modules in the entire configuration.

[0095] ③ Calculate the variable movement distance: Static modules are set as obstacles in the moving module map, and A is calculated through heuristic search. The algorithm calculates the motion paths of modules other than the static module. Let the first module be the first module in the static module. The movement path of each module is The total distance traveled during the metamorphic process is the sum of the number of metamorphic movement steps of all moving modules, i.e.

[0096] (35)

[0097] The length() function represents the length of the path within the parentheses.

[0098] Step 7: Return to Step 5 until the maximum number of iterations is reached. The algorithm ends. After sorting all individuals in the population according to the objective function value and removing duplicates, all non-dominated solutions are the Pareto solutions searched by the algorithm.

[0099] The non-dominated sorting here refers to a solution A and B where A is no worse than (i.e., equal to or better than) B in all optimization objectives, and has at least one objective that is better than B. In this case, A dominates B, and B is dominated by A. If A and B are in a solution set containing multiple solutions, and no solution dominates A, then A is called a non-dominated solution. According to this definition, all solutions are sorted according to their dominance relationships, with non-dominated solutions first and dominated solutions following in order. This process is called non-dominated sorting.

[0100] Example 2

[0101] One embodiment of the present invention provides a modular robot configuration planning system, comprising:

[0102] The separate solution modules are configured to solve for the global optimal solution set for each of the three optimization objectives: minimizing the number of moving modules, minimizing the number of disconnected connecting mechanisms, and minimizing the moving distance.

[0103] The habitat partitioning module is configured to perform hierarchical sequential habitat partitioning on the global optimal solution set of each optimization objective, generating several sub-habitats for each optimization objective, with each sub-habitat containing several variant solutions;

[0104] The feature encoding module is configured to extract common features of the allotropic solutions for each niche and encode them as common chromosome segments.

[0105] The genetic optimization module is configured to: based on common chromosome segments, perform non-dominated sorting iterative optimization within the habitat using a genetic algorithm until the maximum number of iterations is reached, merge the final populations of all habitats, and extract the non-dominated solutions as the optimal solution output.

[0106] Example 3

[0107] One embodiment of the present invention provides a computer program product, including a computer program that, when executed by a processor, implements the modular robot variant planning method.

[0108] Example 4

[0109] In one embodiment of the present invention, a non-transitory computer-readable storage medium is provided for storing computer instructions, which, when executed by a processor, implement the modular robot variant planning method described above.

[0110] Example 5

[0111] One embodiment of the present invention provides an electronic device, including: a processor, a memory, and a computer program; wherein the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory to enable the electronic device to implement the modular robot variant planning method.

[0112] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0113] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0114] While the specific embodiments of the present invention have been described above in conjunction with the accompanying drawings, this is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art without creative effort based on the technical solutions of the present invention are still within the scope of protection of the present invention.

Claims

1. A modular robot configuration planning method, characterized in that, include: For the three optimization objectives of minimizing the number of mobile modules, minimizing the number of disconnected connection mechanisms, and minimizing the travel distance, the global optimal solution set for each optimization objective is solved. The global optimal solution set for each optimization objective is divided into hierarchical sequential habitats, generating several sub-habitats for each optimization objective, with each sub-habitat containing several variant solutions; The process involves hierarchical sequential habitat partitioning of the global optimal solution set for each optimization objective, including partitioning habitat classes and partitioning microhabitats. Habitats include four categories: the set of globally optimal solutions with the fewest movement modules. Habitat class; the globally optimal solution set corresponding to the minimum number of disconnected connecting mechanisms. Habitat category; the globally optimal solution set corresponding to the shortest movement distance Habitat class; Non-optimal solution set correspondence Habitat category; , and Habitat categories all contain multiple subhabitats. The habitat category contains only one small habitat; For each niche, common features of its allotropic solutions are extracted and encoded into common chromosome segments; The extraction of common features from their allosteric solutions and their encoding into common chromosomal segments specifically involves: The microhabitats under the habitat category share the common features of the shape and number of the static modules. The location of the static modules in the target configuration has been determined. The static module numbers are filled into the columns of the row vectors corresponding to the target locations to form common chromosome segments. Microhabitats within habitat categories share the common feature of the distance between metamorphic solutions. The largest common segment of each solution within a habitat is extracted to form a common chromosome segment. Microhabitats within habitat categories share the same alignment point as a common feature, and the module numbers outside the alignment point and the largest rectangular region are considered as common chromosome segments; The largest rectangular area is a rectangular area formed between the mobile module point and its corresponding idle position point; Based on a common chromosome segment, a genetic algorithm is used to perform non-dominated sorting iterative optimization within the habitat until the maximum number of iterations is reached. The final populations of all habitats are then merged, and the non-dominated solution is extracted as the optimal solution for output.

2. The modular robot configuration planning method as described in claim 1, characterized in that, The non-dominated sorting iterative optimization within the habitat using a genetic algorithm is based on an initial population. Chromosome crossover and mutation calculations are performed, and the index values ​​of all modified solutions on the three optimization objectives are independently calculated as the ternary cost of the modified solutions. All modified solutions are then non-dominated sorted, and a new population is selected for the next iteration until the maximum number of iterations is reached.

3. The modular robot configuration planning method as described in claim 2, characterized in that, In the genetic algorithm, each individual in the population represents a feasible solution, and each individual is represented by a chromosome, which is a single chromosome. A 3D row vector consists of two parts: Part One contains numbers 1 through 1. Each element represents the target location of the module; The other part is the alignment point, which is determined by the first... Each element represents the relative position of the initial configuration and the target configuration in the same coordinate system.

4. The modular robot configuration planning method as described in claim 2, characterized in that, The index values ​​for the three optimization objectives are expressed by the following formulas: in, , , These are the number of disconnected connections, the number of moving modules, and the distance the modified module moved. It is the total number of modules in the entire configuration. It represents the number of modules contained in a static component. The number of connections in the initial configuration. Chromosomes The Middle The number of connections for each group, where n is the number of chromosomes. The number of elements in the parentheses, and the length() function represents the length of the path within the parentheses. For the first The movement path of each module.

5. A modular robot configuration planning system, characterized in that, include: The separate solution modules are configured to solve for the global optimal solution set for each of the three optimization objectives: minimizing the number of moving modules, minimizing the number of disconnected connecting mechanisms, and minimizing the moving distance. The habitat partitioning module is configured to perform hierarchical sequential habitat partitioning on the global optimal solution set of each optimization objective, generating several sub-habitats for each optimization objective, with each sub-habitat containing several variant solutions; The process involves hierarchical sequential habitat partitioning of the global optimal solution set for each optimization objective, including partitioning habitat classes and partitioning microhabitats. Habitats include four categories: the set of globally optimal solutions with the fewest movement modules. Habitat class; the globally optimal solution set corresponding to the minimum number of disconnected connecting mechanisms. Habitat category; the globally optimal solution set corresponding to the shortest movement distance Habitat class; Non-optimal solution set correspondence Habitat category; , and Habitat categories all contain multiple subhabitats. The habitat category contains only one small habitat; The feature encoding module is configured to extract common features of the allotropic solutions for each niche and encode them as common chromosome segments. The extraction of common features from their allosteric solutions and their encoding into common chromosomal segments specifically involves: The microhabitats under the habitat category share the common features of the shape and number of the static modules. The location of the static modules in the target configuration has been determined. The static module numbers are filled into the columns of the row vectors corresponding to the target locations to form common chromosome segments. Microhabitats within habitat categories share the common feature of the distance between metamorphic solutions. The largest common segment of each solution within a habitat is extracted to form a common chromosome segment. Microhabitats within habitat categories share the same alignment point as a common feature, and the module numbers outside the alignment point and the largest rectangular region are considered as common chromosome segments; The largest rectangular area is a rectangular area formed between the mobile module point and its corresponding idle position point; The genetic optimization module is configured to: based on common chromosome segments, perform non-dominated sorting iterative optimization within the habitat using a genetic algorithm until the maximum number of iterations is reached, merge the final populations of all habitats, and extract the non-dominated solutions as the optimal solution output.

6. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the modular robot variant planning method according to any one of claims 1-4.

7. A non-transitory computer-readable storage medium, characterized in that, The non-transitory computer-readable storage medium is used to store computer instructions, which, when executed by a processor, implement a modular robot variant planning method as described in any one of claims 1-4.

8. An electronic device, characterized in that, include: The device includes a processor, a memory, and a computer program; wherein the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory to enable the electronic device to implement a modular robot variant planning method as described in any one of claims 1-4.

Images