Rescue robot cooperative control method and system based on disaster area topology analysis

Abstract

The application provides a rescue robot cooperative control method and system based on disaster area topology analysis, and relates to the technical field of cooperative control.The method comprises the following steps: obtaining a three-dimensional scene map of a disaster area, extracting a skeleton map, and constructing a logic controller;obtaining three-dimensional space coordinates and real-time exploration information of each rescue robot, determining a one-dimensional planning strategy and inversely projecting the one-dimensional planning strategy to a three-dimensional space to generate a cooperative control strategy; constructing a Laplace matrix and performing eigenvalue decomposition, identifying rigid subgraphs and flexible subgraphs according to the decomposition, and performing constraint optimization on the cooperative control strategy; determining cooperative logic instructions through strategy decomposition, and cooperatively controlling the rescue robots.The application solves the technical problem that the robot control in the prior art mainly depends on global maps or local sensor information, a large amount of calculation is usually required, especially when multiple robots cooperate, the calculation complexity of path planning increases sharply, and the disaster area rescue task is delayed.

Description

Technical Field

[0001] This invention relates to the field of collaborative control technology, specifically to a collaborative control method and system for rescue robots based on disaster area topology analysis. Background Technology

[0002] In disaster-stricken environments, rescue robots are tasked with rapidly exploring and locating survivors, assessing the environmental conditions of the disaster area, clearing obstacles, and transporting supplies. These tasks typically require robots to be able to cope with complex environmental changes and to maintain stable formation and high efficiency when collaborating with other robots. However, traditional rescue robot control methods have many problems when dealing with complex disaster environments.

[0003] Traditional robot path planning methods mainly rely on global maps or local sensor information, employing grid-based or graph-based path planning algorithms. These algorithms typically require a large amount of computation, especially in complex 3D environments where robots need to consider obstacles, terrain changes, and collaborative relationships with other robots. These computational processes consume significant computing resources and time. In particular, the computational complexity of path planning increases dramatically when multiple robots collaborate. In disaster relief, due to low computational efficiency, robots often cannot adjust their paths in real time to cope with sudden environmental changes, leading to mission delays. In some emergency situations, robots may even be unable to react quickly and execute emergency tasks. Summary of the Invention

[0004] This application provides a collaborative control method and system for rescue robots based on disaster area topology analysis, aiming to solve the technical problem that existing robot control mainly relies on global maps or local sensor information, which usually requires a lot of calculations. Especially when multiple robots cooperate, the computational complexity of path planning increases sharply, leading to delays in disaster area rescue missions.

[0005] The first aspect disclosed in this application provides a collaborative control method for rescue robots based on disaster area topology analysis. The method includes: acquiring a 3D scene map of the target disaster area, extracting a skeleton graph that maintains connectivity, and constructing a logic controller by writing information entropy values, exploration utility relationships, and reward functions; acquiring the 3D spatial coordinates and real-time exploration information of each rescue robot, projecting them onto a 1D skeleton graph within the logic controller, using 1D spacing as the control law, employing information entropy-guided planning, determining a 1D planning strategy, and inversely projecting it onto the 3D space to generate a collaborative control strategy; retrieving the communication graph between each rescue robot, constructing a Laplace matrix and performing eigenvalue decomposition by superimposing the skeleton graph and the communication graph, performing constraint optimization on the collaborative control strategy based on the rigid and flexible subgraphs identified by the decomposition, determining collaborative logic instructions through strategy decomposition, and performing collaborative control of the rescue robots.

[0006] The second aspect of this application discloses a collaborative control system for rescue robots based on disaster area topology analysis. This system is used in the aforementioned collaborative control method for rescue robots based on disaster area topology analysis. The system includes: a logic controller construction module, used to acquire a 3D scene map of the target disaster area, extract a skeleton graph that maintains connectivity, and construct a logic controller by writing information entropy values, exploration utility relationships, and a reward function; a collaborative control strategy generation module, used to acquire the 3D spatial coordinates and real-time exploration information of each rescue robot, project them onto a one-dimensional skeleton graph within the logic controller, use one-dimensional spacing maintenance as the control law, employ information entropy-guided planning, determine a one-dimensional planning strategy, and inversely project it onto 3D space to generate a collaborative control strategy; and a robot collaborative control module, used to retrieve the communication graph between each rescue robot, construct a Laplace matrix by superimposing the skeleton graph and the communication graph, perform eigenvalue decomposition, perform constrained optimization of the collaborative control strategy based on the rigid and flexible subgraphs identified by the decomposition, determine collaborative logic instructions through strategy decomposition, and perform collaborative control of the rescue robots.

[0007] One or more technical solutions provided in this application have at least the following beneficial effects: By acquiring a 3D scene map of the target disaster area and extracting a skeleton graph that maintains connectivity, the complex environment of the disaster area can be effectively simplified, providing a clear and easy-to-operate topology for subsequent robot path planning and reducing redundant operations. By combining information entropy, exploration utility relationships, and reward functions to construct a logic controller, the robot's exploration in the disaster area can be effectively guided, avoiding ineffective repetitive exploration and enabling rapid adjustments to the action plan in the real-time changing disaster environment. Using one-dimensional spacing maintenance as the control law ensures cooperation and formation stability among multiple robots, preventing collisions or formation breakdowns. By applying the control law to the one-dimensional skeleton graph, the robot's movement in the disaster area is optimized. The relative positions of the robots enhance the collective execution efficiency of the robot team. By inversely projecting the robot into three-dimensional space, the generated collaborative control strategy not only ensures the stability of the robot formation but also adapts to complex obstacles and passages in the three-dimensional disaster area environment, ensuring that the actual action path of the robots is feasible and efficient. By overlaying the skeleton graph and the communication graph, and combining the communication graph between robots with the disaster area topology, a joint graph structure is formed, which effectively improves communication efficiency and collaboration accuracy. By constructing a Laplace matrix and using eigenvalue decomposition, rigid subgraphs and flexible subgraphs are identified. This division effectively balances the stability and flexibility of the system, ensuring that the robots can make efficient and accurate collaborations in complex environments.

[0008] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0009] Figure 1 A schematic diagram of the collaborative control method for rescue robots based on disaster area topology analysis provided in this application embodiment.

[0010] Figure 2 A schematic diagram of the collaborative control system structure of a rescue robot based on disaster area topology analysis provided in this application embodiment.

[0011] Figure labeling: Logic controller construction module 10, cooperative control strategy generation module 20, robot cooperative control module 30. Detailed Implementation

[0012] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0013] Example 1, as Figure 1 As shown in the embodiment of this application, a collaborative control method for rescue robots based on disaster area topology analysis is provided, the method comprising: Obtain a 3D scene map of the target disaster area, extract a skeleton diagram that maintains connectivity, and construct a logic controller by writing information entropy values, exploring utility relationships and reward functions.

[0014] A 3D scene map is obtained from the target disaster area. This map reflects the terrain, building structure, and other information of the disaster area, providing a foundation for subsequent path planning and collaborative control. Specifically, 3D information is acquired through technologies such as LiDAR, satellite imagery, and UAV scanning. The 3D scene map is then processed to extract a skeleton graph. The skeleton graph is a simplified topological graph that preserves the main pathways and connectivity of the disaster area. The extraction of the skeleton graph not only retains important structures in the disaster area but also uses the nodes and edges of the skeleton to represent the connection methods within the disaster area, avoiding complex geometric details and facilitating robot path planning.

[0015] To effectively guide the robot's behavior, the logic controller uses information entropy, exploration utility relationships, and a reward function. Information entropy measures the unknown nature of a node; a higher entropy value indicates that the area has not yet been explored, and the robot should prioritize exploring these areas. Exploration utility relationships refer to the robot's utility evaluation during exploration, which is related to factors such as path cost and information gain. The reward function is the reward mechanism for the robot's actions, including the information entropy value collected by the robot, penalties for explored areas, path energy consumption, and other factors. Through this function, the robot can optimize its action strategy.

[0016] The three-dimensional spatial coordinates and real-time exploration information of each rescue robot are acquired and projected onto a one-dimensional skeleton diagram in the logic controller. The one-dimensional spacing is maintained as the control law, and information entropy is used to guide the planning. The one-dimensional planning strategy is determined and inversely projected onto the three-dimensional space to generate a collaborative control strategy.

[0017] Each rescue robot is assigned a three-dimensional coordinate system to represent its current position in the disaster area. Simultaneously, the robots update their exploration information in real time, including changes in the surrounding environment, encountered obstacles, and explored areas. This information is acquired through sensors and transmitted to the control system. The position of each rescue robot is projected onto one dimension of the previously extracted skeleton map, transforming the three-dimensional problem into a one-dimensional one. The projection process uses arc-length coordinates, allowing the robot's position to be expressed as the cumulative arc distance to the starting point of the skeleton. Maintaining the spacing between robot positions in the one-dimensional map serves as a control law, ensuring cooperation and positional relationships between robots. The logic controller uses an information entropy-guided path planning method, prioritizing each node based on its information entropy and selecting the path with the highest information gain for exploration. The planned strategy is derived from the one-dimensional projection map and then inversely projected back into three-dimensional space to generate the robot's cooperative control strategy. This approach ensures that the robots perform the most effective exploration tasks in the disaster area while maintaining formation and mutual cooperation.

[0018] The communication graph between each rescue robot is retrieved. By overlaying the skeleton graph and the communication graph, a Laplace matrix is ​​constructed and eigenvalue decomposition is performed. Based on the rigid and flexible subgraphs identified by the decomposition, the cooperative control strategy is constrained and optimized. Cooperative logic instructions are determined through strategy decomposition to perform cooperative control of the rescue robots.

[0019] In a multi-robot system, each robot has the ability to communicate with other robots. In this step, a communication graph between the robots is obtained through the communication system, where nodes represent robots and edges represent communication connections between robots. The skeleton graph is overlaid with the communication graph to construct a new joint graph structure. The nodes in this joint graph not only represent robots but also include communication links between robots and path constraints based on the skeleton graph. This means that not only the communication capabilities between robots affect collaborative control but also their physical location and accessibility in the disaster area.

[0020] The Laplacian matrix is ​​an important matrix used to describe graph structure. By calculating the Laplacian matrix of the joint graph and performing eigenvalue decomposition, rigid and flexible subgraphs can be identified: rigid subgraphs contain tightly connected clusters of robots, with relatively strong communication and cooperation among them. A larger eigenvalue indicates that this part of the subgraph has strong connectivity and high cooperative ability. Flexible subgraphs are looser subsets of robots, with smaller eigenvalues, indicating that the connectivity between these robots is weaker and more coordination and control are needed to maintain effective cooperation.

[0021] Based on the decomposition results of rigid and flexible subgraphs, the cooperative control strategy is optimized. Robots on rigid subgraphs maintain consistency through close cooperation, while those on flexible subgraphs require constraint optimization, involving position adjustments and communication delay compensation. Finally, through strategy decomposition, different cooperative logic instructions are generated. These instructions are transmitted to each rescue robot, and each instruction is optimized according to the robot's task and its subgraph, ensuring that all robots work collaboratively to effectively complete the disaster relief mission.

[0022] Furthermore, the skeleton graph defines a set of nodes with the region center and key branch points, and a set of edges with the centerline of walkable paths. The edges are labeled with a width attribute representing the minimum width. Based on the skeleton graph, a logic controller is built using a collaborative planning method guided by information entropy as the training logic.

[0023] The set of nodes in the skeleton diagram is defined by the regional center and key branch points. The regional center represents the key area within the disaster area. It is a relatively stable and easy-to-find central point, usually the center of gravity of the disaster area. Key branch points refer to important intersections or forks in the disaster area, such as road junctions or building connection points, which can directly affect the subsequent path selection.

[0024] The set of edges in the skeleton graph is defined by the centerlines of walkable pathways within the disaster area. These centerlines refer to traversable areas within the disaster area, such as roads and passageways. Each centerline of a walkable pathway becomes an edge in the skeleton graph. Each edge not only represents a connection between pathways but also indicates the minimum width attribute of the pathway. The width attribute characterizes the pathway's capacity and directly affects the robot's spatial perception and motion planning when traversing that area. If the pathway is too narrow, it will affect the robot's ability to pass through.

[0025] The skeleton diagram, serving as the foundation of the entire collaborative control system, provides the basic topological structure within the disaster area, ensuring the robot can navigate effectively in this environment. Information entropy-guided collaborative planning, a path planning method based on information gain, helps the robot find the most information-valuable path in the disaster area. Information entropy represents the uncertainty of a region or node; higher information entropy means less information in that region. The robot prioritizes exploring areas with higher information entropy to maximize the acquisition of new information. Through information entropy-guided collaborative planning, a logic controller is built. This controller directs the robot's actions based on the disaster area's topology, information entropy values, and exploration utility, ensuring the robot can make decisions based on real-time information and execute the optimal collaborative exploration task.

[0026] Furthermore, an information entropy value is defined for each node in the node set, wherein the information entropy value is positively correlated with the node connectivity, the number of unexplored branches, and the rescue probability coefficient; an exploration utility relationship is defined, wherein the exploration utility relationship is defined by information gain and path cost, wherein the information gain is the amount of new information obtained after reaching a node, including the sum of the information entropy of the corresponding node and its neighboring unexplored edges; a reward function for the robot is defined, wherein the reward function is defined by the individual collected information entropy - the penalty term for explored areas - the path energy consumption; and the information entropy value, the exploration utility relationship, and the reward function are written into the logic controller.

[0027] An information entropy value is defined for each node. Node connectivity reflects the degree to which the node is connected to other nodes; higher connectivity indicates greater importance of the node in the skeleton graph and lower information entropy. The number of unexplored branches represents the number of unexplored paths among the branches connected to the node; more branches mean more unexplored paths and a higher information entropy value. The rescue probability coefficient represents the rescue success rate of the area where the node is located; areas with high success rates have lower information entropy, and areas with low success rates have higher information entropy. The positive correlation of these factors indicates that the exploration priority of a node is related not only to its own connectivity but also to the number of unexplored branches and the rescue availability of the area.

[0028] Exploration utility relations are used to measure the efficiency of a robot in performing exploration tasks. Information gain represents the amount of new information the robot acquires after reaching a node. Information gain includes the information entropy of the node itself and the sum of the information entropy of its adjacent unexplored edges. High information gain means that the node contributes significantly to the overall task and is a region that the robot should prioritize exploring. Path cost refers to the cost required for the robot to travel from the current node to the target node, including time consumption, energy consumption, and path complexity. Routes with lower path costs are preferred. By considering these factors, exploration utility relations help the robot select the optimal path within a disaster area, ensuring that the robot explores in the most efficient way while avoiding unnecessary energy waste and time delays.

[0029] When robots perform tasks, a reward function is used to evaluate the value of each action and drive the robot to choose the optimal path. Each robot collects information about its environment during exploration tasks; an increase in information entropy reflects the new information gathered. The collection of information entropy helps improve the overall system's understanding of the disaster area; therefore, collecting more information entropy is a reward for robot actions. If a robot enters an already explored area, it means repetitive work with no benefit, so a penalty is imposed. This penalty reduces the robot's tendency to explore known areas, prompting it to explore unknown areas, thereby improving task efficiency. Path energy consumption refers to the energy or time consumed by the robot during task execution. Lower path energy consumption results in higher task completion efficiency; therefore, paths with lower energy consumption yield higher rewards. Considering these factors, the reward function can be expressed as: reward value equals information entropy minus the penalty for explored areas minus path energy consumption. Through this reward function, the robot can autonomously choose and optimize its behavior during task execution, achieving optimal exploration and collaboration results.

[0030] By integrating all control elements, including information entropy values, exploration utility relationships, and reward functions, into the logic controller, a unified decision-making framework is formed. This allows the logic controller to make real-time decisions, providing guidance for the collaborative control of robots and ensuring that each robot explores and collaborates according to the optimal strategy.

[0031] Furthermore, generating collaborative control strategies includes: The three-dimensional spatial coordinates are projected onto a one-dimensional skeleton graph through nearest neighbor search. The projection method is as follows: the position is represented by arc length coordinates, which are the cumulative arc lengths from the projection point to the starting point of the skeleton; the formation is represented by the relative distance sequence on the edge line; the information entropy value of each node is initialized according to the real-time exploration information; robot cooperative planning is performed in the one-dimensional skeleton graph to determine the cooperative control strategy.

[0032] The nearest neighbor search method projects 3D spatial coordinates onto a 1D skeleton graph. Nearest neighbor search involves finding the node in the skeleton graph closest to the robot's current position and projecting it onto that node. Arc length coordinates represent a position in space as the cumulative arc length from the starting point (the starting node of the skeleton graph) to the current position. The position of each node is determined by its cumulative arc length from the starting point of the skeleton graph. In this way, the robot can accurately represent its position in the disaster area in one-dimensional space. After projection, the robots maintain a certain formation based on their relative positions. The relative distance sequence along the edges ensures that the robots maintain a certain formation during collaborative tasks, facilitating effective cooperation and coordination.

[0033] Each robot continuously acquires real-time exploration information during the exploration process, including changes in its surrounding environment, encountered obstacles, and explored areas. Based on this information, the controller initializes the information entropy value for each node. The purpose of information entropy initialization is to assign an initial value to each node, representing the unknown nature of that area on the disaster map. A higher information entropy value means that the area has not yet been explored and has a higher priority. The robots perform collaborative planning through a logic controller on a one-dimensional skeleton graph. The goal of collaborative planning is to enable multiple robots to cooperate effectively during exploration, avoiding conflicts and redundant exploration, while maximizing efficiency. The planning process combines information entropy values, exploration utility relationships, and reward functions to help the robots determine their movement paths between different nodes, ensuring that the robots can optimize their decisions based on current exploration information. Finally, a collaborative control strategy is generated based on the planning results. This strategy guides each robot on how to move and cooperate in the disaster area and dynamically adjusts its path planning based on real-time exploration information.

[0034] Furthermore, performing robot cooperative planning in a one-dimensional skeleton graph includes: Using one-dimensional spacing as the control law, and based on the exploration utility relationship and reward function, planning guided by information entropy is performed to determine a one-dimensional planning strategy; the one-dimensional planning strategy is then back-projected into three-dimensional space to determine the cooperative control strategy.

[0035] In collaborative robot control, a one-dimensional spacing maintenance control law is used to ensure that multiple robots can effectively maintain their formation and avoid collisions or excessive dispersion during exploration. One-dimensional spacing maintenance refers to maintaining a certain distance between robots on a one-dimensional skeleton diagram. This ensures the neatness and compactness of the robot formation. The position on the one-dimensional skeleton diagram determines the robot's location and relative formation within the disaster area. Maintaining a certain distance helps ensure coordination between robots and the efficiency of the exploration task.

[0036] To guide the robot to explore the optimal path, the robot's path planning is optimized by exploring utility relationships and reward functions. By using information entropy as a guide, the robot can prioritize exploring areas with higher information entropy, thereby maximizing the effect of information collection. Through the above mechanism, a one-dimensional planning strategy is generated.

[0037] Inverse projection refers to transforming the planning results on a one-dimensional skeleton diagram back into a three-dimensional path and control strategy. This is because the actual operation of the robot takes place in three-dimensional space. Therefore, it is necessary to interface the one-dimensional control strategy with the actual environment to ensure that the planned path is feasible in the three-dimensional environment. After inverse projection, the resulting cooperative control strategy can clearly define the specific action steps of each robot, ensuring that the robots can work effectively and collaboratively in the disaster area and dynamically adjust according to the real-time situation.

[0038] Furthermore, by overlaying the skeleton graph and the communication graph, a Laplacian matrix is ​​constructed and eigenvalue decomposition is performed, including: Obtain the communication graph of robots within the formation; define a joint graph structure by overlaying the skeleton graph and the communication graph, wherein in the joint graph structure, nodes are robots, and edges include communication links between robots and path constraints between projection points of adjacent robots on the skeleton graph; based on the joint graph structure, identify the rigid and flexible subgraphs of the formation by constructing a Laplacian matrix and performing eigenvalue decomposition.

[0039] Each robot can communicate with other robots when performing a task. The communication graph is a graph composed of the robots' communication capabilities and communication links, where each node represents a robot and the edges represent the communication connections between robots.

[0040] By combining the skeleton graph and the communication graph, a new joint graph structure is formed. The nodes in the joint graph represent various rescue robots, and the edges include not only the communication links between robots, but also the path constraints between adjacent projection points of robots in the skeleton graph. The path constraints represent the physical paths that robots need to take, ensuring that they can coordinate according to the prescribed formation and path.

[0041] The Laplacian matrix is ​​a tool in graph theory that reflects the structural characteristics of a graph. Based on the structure of a joint graph, the collaborative relationships between robots can be analyzed by constructing the Laplacian matrix. The degree matrix is ​​a diagonal matrix where the diagonal elements represent the number of edges connected to each node (i.e., each robot) in the joint graph. The adjacency matrix is ​​a square matrix where the elements represent whether there are communication links between robots and the weight of those links. Through eigenvalue decomposition, based on the connectivity between robots, it is possible to identify which robots belong to rigid subgraphs and which belong to flexible subgraphs. Rigid subgraphs are tightly connected parts of the graph where robots have strong connections and must maintain fixed relative positions to ensure the entire formation does not fall apart; these robots play a core role in the formation. Flexible subgraphs are located at the edges of the graph or in greater contact with the environment, with weaker connectivity, allowing these robots to flexibly adjust their positions within a certain range to cope with obstacles or narrow passages. When the environment changes, robots in rigid subgraphs can remain stable, while robots in flexible subgraphs can flexibly adjust their positions to enhance the adaptability of the entire team in complex and changing environments.

[0042] Furthermore, for the joint graph structure, a degree matrix and an adjacency matrix are constructed, and the Laplacian matrix is ​​obtained by subtracting the matrices. The degree matrix is ​​a diagonal matrix, where the diagonal elements correspond to the number of connecting edges for each rescue robot node in the joint graph structure, and the off-diagonal elements are zero. The adjacency matrix is ​​a square matrix, where the rows and columns correspond to the rescue robot nodes. When there are connecting edges between rescue robot nodes, the matrix elements are the edge weights, and when there are no connecting edges, the matrix elements are zero.

[0043] In graph theory, the degree matrix and adjacency matrix are two matrices that describe the structure of a graph, and they provide the foundation for constructing the Laplacian matrix. The Laplacian matrix is ​​obtained by subtracting the degree matrix from the adjacency matrix, and it describes the connectivity and structural properties of the graph.

[0044] The degree matrix is ​​a diagonal matrix. The diagonal elements of the matrix represent the degree of each node (in this case, each robot) in the joint graph, which is the number of connections that node has with other nodes. The off-diagonal elements of the degree matrix are zero because the degree matrix only represents the number of connections for each node and does not involve the direct connection relationships between nodes. The connections include not only the communication links between robots, but also the path constraints between the robot and adjacent projection points in the skeleton graph.

[0045] An adjacency matrix is ​​a square matrix that describes the connection relationships between nodes in a graph. In an adjacency matrix, if two nodes are connected by an edge, the corresponding position in the matrix will be filled with the weight of the edge. If two nodes are not connected by an edge, the corresponding position in the matrix will be zero. The adjacency matrix provides comprehensive information about how nodes in the graph are directly connected.

[0046] Furthermore, the cooperative control strategy is constrained and optimized, and cooperative logic instructions are determined through strategy decomposition to achieve cooperative control of the rescue robot, including: Using the rigid and flexible subgraphs as elastic constraints, the cooperative control strategy is optimized to generate multi-threaded cooperative logic instructions; these instructions are then sent to the corresponding rescue robots to execute cooperative drive control.

[0047] Elastic constraints refer to how to optimize the cooperative control strategy to adapt to the different needs of rigid and flexible subgraphs. Robots in rigid subgraphs require higher constraints and stability, while robots in flexible subgraphs have more degrees of freedom to adjust. In this step, the optimization objective is to generate a reasonable cooperative control strategy that adapts to different environmental changes through the elastic constraints of rigid and flexible subgraphs. The optimization process imposes strong positioning constraints on robots in rigid subgraphs to maintain formation integrity, while providing greater flexibility for robots in flexible subgraphs, allowing them to adaptively adjust their positions when the environment changes. The optimized cooperative control strategy generates corresponding multi-threaded cooperative logic instructions, which are then issued to different robots to enable them to cooperate efficiently when executing tasks. Multi-threaded instructions mean that each robot will execute different operations or tasks in parallel, ensuring that the entire team can work synchronously and avoiding conflicts or information delays between robots.

[0048] The generated multi-threaded collaborative logic instructions are transmitted to each corresponding rescue robot. Each robot executes specific task operations according to the instructions and synchronizes its actions in the collaborative control system to ensure effective cooperation and avoidance of conflict. The goal of collaborative drive control is to ensure that robots can cooperate effectively in the disaster area, not only avoiding obstacles but also ensuring the stability of the formation and the effective completion of the mission.

[0049] Furthermore, the method also includes: For key branch points, rescue robots are deployed based on the logical rules of splitting and merging. The logical rules are deployed in one dimension with the branching direction of the key branch point as the reference. When the rescue robot passes through any key branch point, motion control commands based on the logical rules are generated.

[0050] In disaster-stricken environments, key branching points, such as road intersections and path forks in rubble, require robots to make decisions based on different directions. Logical rules for splitting and merging are strategies for handling these situations. Splitting refers to the need for multiple robots to separate and explore different paths when approaching a branching point. Logical rules ensure that robots are rationally distributed across various paths to avoid congestion. Merging refers to the need for robots to merge back into the main formation at a certain point after exploring different paths. Logical rules ensure that robots can regroup at the appropriate time.

[0051] By using the branching direction of key branch points as the basis for deploying one-dimensional logical rules, the robot will decide whether to split or merge based on the branching direction. Furthermore, each robot makes path selection and decisions based on its position and task objective. This effectively avoids collisions between robots in narrow passages or complex environments, while improving task execution efficiency.

[0052] By employing dimensionality reduction control, path planning is transformed from a three-dimensional problem into a one-dimensional problem, reducing the amount of computation. This is especially beneficial when multiple robots are performing path planning simultaneously. Controlling the path on a one-dimensional skeleton diagram reduces the amount of real-time communication between robots, as control information only needs to be transmitted in one-dimensional space, rather than transmitting three-dimensional coordinate data every moment.

[0053] As the robot traverses critical branch points, motion control commands are generated based on splitting and merging logic rules. These commands ensure the robot can correctly split or merge at these points and continue performing its task. The motion control commands are dynamically adjusted according to real-time environmental changes and the tasks the robot is performing, ensuring the robot can flexibly respond to various changes in the disaster area environment.

[0054] Example 2, based on the same inventive concept as the disaster area topology analysis-based collaborative control method for rescue robots in the previous examples, such as... Figure 2 As shown in the figure, this application provides a collaborative control system for rescue robots based on disaster area topology analysis. The system includes: The logic controller construction module 10 is used to acquire a 3D scene map of the target disaster area, extract a skeleton graph that maintains connectivity, and construct a logic controller by writing information entropy values, exploring utility relationships, and reward functions. The cooperative control strategy generation module 20 is used to acquire the 3D spatial coordinates and real-time exploration information of each rescue robot, project them onto the 1D skeleton graph within the logic controller, use 1D spacing as the control law, adopt information entropy-guided planning, determine the 1D planning strategy, and inversely project it onto the 3D space to generate a cooperative control strategy. The robot cooperative control module 30 is used to retrieve the communication graph between each rescue robot, construct a Laplace matrix by superimposing the skeleton graph and the communication graph, and perform eigenvalue decomposition. Based on the rigid and flexible subgraphs identified by the decomposition, the cooperative control strategy is constrained and optimized. Cooperative logic instructions are determined through strategy decomposition to perform cooperative control of the rescue robots.

[0055] Furthermore, the skeleton graph defines a set of nodes with the region center and key branch points, and a set of edges with the centerline of walkable paths. The edges are labeled with a width attribute representing the minimum width. Based on the skeleton graph, a logic controller is built using a collaborative planning method guided by information entropy as the training logic.

[0056] Furthermore, an information entropy value is defined for each node in the node set, wherein the information entropy value is positively correlated with the node connectivity, the number of unexplored branches, and the rescue probability coefficient; an exploration utility relationship is defined, wherein the exploration utility relationship is defined by information gain and path cost, wherein the information gain is the amount of new information obtained after reaching a node, including the sum of the information entropy of the corresponding node and its neighboring unexplored edges; a reward function for the robot is defined, wherein the reward function is defined by the individual collected information entropy - the penalty term for explored areas - the path energy consumption; and the information entropy value, the exploration utility relationship, and the reward function are written into the logic controller.

[0057] Furthermore, the collaborative control strategy generation module 20 is used to perform the following operation steps: The three-dimensional spatial coordinates are projected onto a one-dimensional skeleton graph through nearest neighbor search. The projection method is as follows: the position is represented by arc length coordinates, which are the cumulative arc lengths from the projection point to the starting point of the skeleton; the formation is represented by the relative distance sequence on the edge line; the information entropy value of each node is initialized according to the real-time exploration information; robot cooperative planning is performed in the one-dimensional skeleton graph to determine the cooperative control strategy.

[0058] Furthermore, the collaborative control strategy generation module 20 is used to perform the following operation steps: Using one-dimensional spacing as the control law, and based on the exploration utility relationship and reward function, planning guided by information entropy is performed to determine a one-dimensional planning strategy; the one-dimensional planning strategy is then back-projected into three-dimensional space to determine the cooperative control strategy.

[0059] Furthermore, the robot collaborative control module 30 is used to perform the following operation steps: Obtain the communication graph of robots within the formation; define a joint graph structure by overlaying the skeleton graph and the communication graph, wherein in the joint graph structure, nodes are robots, and edges include communication links between robots and path constraints between projection points of adjacent robots on the skeleton graph; based on the joint graph structure, identify the rigid and flexible subgraphs of the formation by constructing a Laplacian matrix and performing eigenvalue decomposition.

[0060] Furthermore, for the joint graph structure, a degree matrix and an adjacency matrix are constructed, and the Laplacian matrix is ​​obtained by subtracting the matrices. The degree matrix is ​​a diagonal matrix, where the diagonal elements correspond to the number of connecting edges for each rescue robot node in the joint graph structure, and the off-diagonal elements are zero. The adjacency matrix is ​​a square matrix, where the rows and columns correspond to the rescue robot nodes. When there are connecting edges between rescue robot nodes, the matrix elements are the edge weights, and when there are no connecting edges, the matrix elements are zero.

[0061] Furthermore, the robot collaborative control module 30 is used to perform the following operation steps: Using the rigid and flexible subgraphs as elastic constraints, the cooperative control strategy is optimized to generate multi-threaded cooperative logic instructions; these instructions are then sent to the corresponding rescue robots to execute cooperative drive control.

[0062] Furthermore, the robot collaborative control module 30 is used to perform the following operation steps: For key branch points, rescue robots are deployed based on the logical rules of splitting and merging. The logical rules are deployed in one dimension with the branching direction of the key branch point as the reference. When the rescue robot passes through any key branch point, motion control commands based on the logical rules are generated.

[0063] Through the foregoing detailed description of the collaborative control method for rescue robots based on disaster area topology analysis, those skilled in the art can clearly understand the collaborative control system for rescue robots based on disaster area topology analysis in this embodiment. Since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and relevant parts can be referred to in the method section.

[0064] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A collaborative control method for rescue robots based on disaster area topology analysis, characterized in that, The method includes: Obtain a 3D scene map of the target disaster area, extract a skeleton diagram that maintains connectivity, and construct a logic controller by writing information entropy values, exploring utility relationships and reward functions; The three-dimensional spatial coordinates and real-time exploration information of each rescue robot are obtained and projected onto a one-dimensional skeleton diagram in the logic controller. The one-dimensional spacing is maintained as the control law. Information entropy is used to guide the planning, determine the one-dimensional planning strategy, and inversely project it onto the three-dimensional space to generate a cooperative control strategy. The communication graph between each rescue robot is retrieved. By overlaying the skeleton graph and the communication graph, a Laplace matrix is ​​constructed and eigenvalue decomposition is performed. Based on the rigid and flexible subgraphs identified by the decomposition, the cooperative control strategy is constrained and optimized. Cooperative logic instructions are determined through strategy decomposition to perform cooperative control of the rescue robots.

2. The method of claim 1, wherein, The skeleton diagram defines a set of nodes with the region center and key branch points, and defines a set of edges with the centerline of walkable paths. The edges are labeled with a width attribute that represents the minimum width. Based on the skeleton diagram, and using a collaborative planning approach guided by information entropy as the training logic, a logic controller is built.

3. The collaborative control method for rescue robots based on disaster area topology analysis as described in claim 2, characterized in that, An information entropy value is defined for each node in the node set, wherein the information entropy value is positively correlated with the node connectivity, the number of unexplored branches, and the probability coefficient of rescue. Define the exploration utility relationship, wherein the exploration utility relationship is defined by information gain and path cost, and the information gain is the amount of new information obtained after reaching a node, including the sum of the information entropy of the corresponding node and the neighboring unexplored edges; Define the robot's reward function, wherein the reward function is defined by the entropy of personally collected information - the penalty term for explored areas - the path energy consumption; The information entropy value, the exploration utility relationship, and the reward function are written into the logic controller.

4. The collaborative control method for rescue robots based on disaster area topology analysis as described in claim 3, characterized in that, Generate collaborative control strategies, including: The three-dimensional spatial coordinates are projected onto a one-dimensional skeleton graph through nearest neighbor search. The projection method is as follows: the position is represented by arc length coordinates, which are the cumulative arc lengths from the projection point to the starting point of the skeleton, and the formation is represented by the relative distance sequence on the edge line. Based on the real-time exploration information, the information entropy value of each node is initialized, and robot cooperative planning is performed in a one-dimensional skeleton graph to determine the cooperative control strategy.

5. The collaborative control method for rescue robots based on disaster area topology analysis as described in claim 4, characterized in that, Performing cooperative robot planning in a one-dimensional skeleton graph includes: Using one-dimensional spacing as the control law, and based on the aforementioned exploration utility relationship and reward function, planning guided by information entropy is performed to determine the one-dimensional planning strategy. The one-dimensional planning strategy is back-projected into three-dimensional space to determine the cooperative control strategy.

6. The collaborative control method for rescue robots based on disaster area topology analysis as described in claim 1, characterized in that, By overlaying the skeleton graph and the communication graph, a Laplacian matrix is ​​constructed and eigenvalue decomposition is performed, including: Obtain the communication graph of robots within the formation; By superimposing the skeleton graph and the communication graph, a joint graph structure is defined, wherein the nodes in the joint graph structure are robots, and the edges include communication links between robots and path constraints between adjacent robot projection points on the skeleton graph. Based on the joint graph structure, rigid and flexible subgraphs of the formation are identified by constructing a Laplacian matrix and performing eigenvalue decomposition.

7. The collaborative control method for rescue robots based on disaster area topology analysis as described in claim 6, characterized in that, For the joint graph structure, construct the degree matrix and the adjacency matrix, and obtain the Laplace matrix by subtracting the matrices; The degree matrix is ​​a diagonal matrix, where the diagonal elements correspond to the number of connecting edges of each rescue robot node in the joint graph structure, and the off-diagonal elements are zero. The adjacency matrix is ​​a square matrix, with rows and columns corresponding to rescue robot nodes. When there are connecting edges between rescue robot nodes, the matrix elements are edge weights; when there are no connecting edges, the matrix elements are zero.

8. The collaborative control method for rescue robots based on disaster area topology analysis as described in claim 1, characterized in that, The cooperative control strategy is constrained and optimized. Cooperative logic instructions are determined through strategy decomposition to perform cooperative control of the rescue robot, including: Using the rigid and flexible subgraphs as elastic constraints, the cooperative control strategy is optimized to generate multi-threaded cooperative logic instructions. The multi-threaded collaborative logic instructions are sent to the corresponding rescue robot to execute collaborative drive control.

9. The collaborative control method for rescue robots based on disaster area topology analysis as described in claim 1, characterized in that, The method further includes: For key branch points, rescue robots are deployed based on the logical rules of splitting and merging, wherein the one-dimensional logical rules are deployed based on the branching direction of the key branch points. When the rescue robot passes through any key branch point, it generates motion control commands based on the aforementioned logical rules.

10. A collaborative control system for rescue robots based on disaster area topology analysis, characterized in that, The system is used to implement the disaster area topology analysis-based collaborative control method for rescue robots according to any one of claims 1-9, the system comprising: The logic controller construction module is used to acquire a 3D scene map of the target disaster area, extract a skeleton graph that maintains connectivity, and construct the logic controller by writing information entropy values, exploring utility relationships and reward functions. The collaborative control strategy generation module is used to acquire the three-dimensional spatial coordinates and real-time exploration information of each rescue robot, project them onto the one-dimensional skeleton diagram in the logic controller, use one-dimensional spacing as the control law, adopt information entropy to guide planning, determine the one-dimensional planning strategy and back-project it to three-dimensional space to generate a collaborative control strategy. The robot collaborative control module is used to retrieve the communication graph between each rescue robot, construct a Laplace matrix by superimposing the skeleton graph and the communication graph, and perform eigenvalue decomposition. Based on the rigid and flexible subgraphs identified by the decomposition, the collaborative control strategy is constrained and optimized. The collaborative logic instructions are determined through strategy decomposition to perform collaborative control of the rescue robots.

Images