Heterogeneous robot cooperation method and system

Abstract

The invention relates to the technical field of heterogeneous robot cooperation, and provides a heterogeneous robot cooperation method and system, and the method comprises the steps: constructing a multi-protocol adaptive communication adaptation layer, achieving the heterogeneous robot data interaction of CAN, Ethernet / IP, ROS and Modbus protocols based on an intelligent protocol conversion unit, controlling the communication delay [tau] to be smaller than or equal to 45ms through an improved delay compensation model, and carrying out the data interaction of a heterogeneous robot. Real-time state parameters and dynamic capability factors of the heterogeneous robots are collected, the state parameters comprise the residual electric quantity S, the current load L, the movement speed v and the equipment health degree H, and the dynamic capability factors comprise the movement adaptation coefficient M and the operation precision coefficient P. Through the multi-protocol adaptive communication adaptation layer and the improved delay compensation model, low-delay data transmission is realized, the link stability is remarkably enhanced, the communication adaptation performance is further improved in a breakthrough manner, seamless interaction of multiple mainstream protocols is supported, and the problems of poor protocol compatibility and insufficient communication reliability are effectively solved.

Description

Technical Field

[0001] This invention relates to the field of heterogeneous robot collaboration technology, specifically to a heterogeneous robot collaboration method and system. Background Technology

[0002] Heterogeneous robot collaboration technology, as a core supporting technology in fields such as intelligent manufacturing, logistics warehousing, and emergency rescue, can integrate the advantages of robots with different functions and specifications to achieve complex tasks that are difficult for a single robot to complete. Its application scenarios are constantly expanding, and market demand continues to grow. With the development of Industry 4.0 and intelligent equipment technology, the requirements for the real-time performance, reliability, and high precision of heterogeneous robot collaboration are becoming increasingly stringent, especially the collaborative performance under multi-scenario switching and dynamic task loads, which has become a key direction for technological breakthroughs.

[0003] Existing heterogeneous robot collaboration technologies face several pressing issues: First, poor communication protocol compatibility, with most systems supporting only a few fixed protocols, resulting in high data interaction latency and insufficient stability, making it difficult to adapt to the collaborative needs of multiple robots in complex environments. Second, static task allocation mechanisms, often based on preset rules or single optimization objectives, fail to adequately consider real-time adaptation to changes in robot dynamic capabilities and task requirements, leading to uneven resource consumption and low task completion efficiency. Third, limited collaborative positioning accuracy; traditional multi-source fusion positioning methods lack dynamic calibration based on relative position constraints between robots, making them prone to positioning drift in complex environments, affecting collaboration accuracy. Fourth, weak fault response and reconfiguration capabilities; limited fault detection methods and time-consuming task redistribution make it difficult to ensure continuous task execution. Fifth, a lack of a global optimization scheduling mechanism; communication, positioning, and task allocation modules operate independently, failing to create a collaborative optimization effect, thus limiting overall collaboration efficiency.

[0004] Current technical solutions have failed to effectively address the aforementioned comprehensive issues, particularly in areas such as multi-protocol adaptive compatibility, deep integration of dynamic capability perception and task allocation, collaborative optimization of high-precision positioning and low-latency communication, and rapid fault reconstruction and task continuity assurance. Existing technologies either focus on optimizing the performance of a single module or lack the ability to adapt to dynamic environments and task changes. This results in the reliability and efficiency of heterogeneous robot collaboration in complex scenarios failing to meet practical application requirements, thus limiting its large-scale application in key areas such as high-end manufacturing and emergency rescue. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a heterogeneous robot collaboration method and system, which solves the problem of poor communication protocol compatibility.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a heterogeneous robot collaboration method and system, comprising the following steps: Step 1: Construct a multi-protocol adaptive communication adaptation layer. Based on an intelligent protocol conversion unit, realize heterogeneous robot data interaction using CAN, Ethernet / IP, ROS, and Modbus protocols. Control communication latency through an improved latency compensation model. ≤45ms; Step 2: Collect real-time status parameters and dynamic capability factors of each heterogeneous robot. The status parameters include remaining power S, current load L, movement speed v, and device health H. The dynamic capability factors include movement adaptation coefficient M and operation accuracy coefficient P. Step 3: Based on the improved particle swarm optimization-auction fusion algorithm, construct the optimization objective function for the task allocation unit: in, , , ,and ; This represents the total time required to complete the task. The variance of resource consumption for each robot, This function determines the task-robot fit and enables dynamic task allocation. Step 4: Employ a vision-IMU-LiDAR-ultrasound multi-source fusion positioning method, combined with relative position constraints and dynamic calibration mechanisms among heterogeneous robots; Step 5: Monitor the robot's operating status in real time. When a fault occurs, trigger the intelligent dynamic reconfiguration unit to complete the task reassignment of the faulty robot within 80ms.

[0007] Preferably, the calculation formula for the improved delay compensation model is as follows: ,in, For communication distance, For data bandwidth, This represents the protocol complexity coefficient; the parameter value is... =0.018ms / m, =0.0008ms / MB, =0.5ms; transmission delay is offset by a two-stage algorithm of pre-compensation-real-time calibration.

[0008] Preferably, the weighted calculation formula for the equipment health H is: ,in, =0.45, =0.25, =0.15, =0.15, and =1; For hardware health status, This refers to the software's running status. For environmental adaptation status, Task fatigue level.

[0009] Preferably, the key parameters of the improved particle swarm-auction fusion algorithm are set as follows: Inertia weight calculation formula: ,in, =0.95, =0.35; This represents the current iteration number. This represents the maximum number of iterations. Learning factors: c1=2.1, c2=2.1; Auction bidding threshold: 0.75; Particle population size: 40-60; Algorithm convergence accuracy ≤5e-4.

[0010] Preferably, in step four, the multi-source fusion localization uses an adaptive extended Kalman filter, with the core equations and parameters as follows: State equation: Observation equation: ,in, The position-attitude-velocity state vector, To control the input, For process noise, For observation noise; initial process noise covariance matrix: Q=diag([8e-5,8e-5,8e-5,5e-7,5e-7,5e-7]); initial observation noise covariance matrix: R=diag([8e-4,8e-4,8e-4]); Q and R are dynamically updated every 10 sampling periods to adapt to environmental changes.

[0011] A heterogeneous robot collaborative system, comprising: Communication adaptation module: includes intelligent protocol parsing unit, dynamic data format conversion unit, and improved delay compensation unit, controlling communication delay τ≤45ms; Status and Capability Perception Module: Includes a high-precision power detection unit, a load dynamic monitoring unit, a speed acquisition unit, a multi-dimensional health assessment unit, and a capability factor calculation unit, which collects status parameters and dynamic capability factors; Intelligent task allocation module: includes algorithm operation unit, parameter input unit, allocation execution unit, and bid evaluation unit, based on improved particle swarm optimization-auction fusion algorithm and objective function. Achieve dynamic allocation; High-precision collaborative positioning module: includes multi-source data acquisition unit, adaptive fusion filtering unit, dynamic positioning calibration unit, and relative position constraint unit, achieving positioning accuracy ≤1.5cm; Intelligent fault reconstruction module: includes dual-mode fault detection unit, rapid task redistribution unit, status feedback unit, and task continuity guarantee unit, with a fault response time ≤80ms; Central control module: includes instruction generation unit, data storage unit, logic judgment unit, and global optimization scheduling unit, and communicates bidirectionally with the above modules and coordinates them in an overall manner; Adaptive communication topology management unit: includes neighbor discovery unit, link quality assessment unit, route optimization unit, and topology reconstruction unit, to build dynamic Mesh network topology.

[0012] Preferably, the judgment rule of the dual-mode fault detection unit is as follows: a fault is determined when a certain state parameter meets any of the following conditions: the monitoring threshold of claim 3 is exceeded for four consecutive sampling periods; the absolute value of the slope of the state parameter change trend is ≥0.7; the abrupt change amplitude of the state parameter is ≥20%.

[0013] Preferably, the link quality assessment unit calculates the link quality indicator using the following formula: ,in, For receiving signal strength indication, For bit error rate, Transmission delay; weighting factor values =0.4, =0.35, γ=0.25; when The topology reconstruction algorithm is triggered when the value is less than 0.6.

[0014] Preferably, it also includes a simulation verification and performance evaluation module, which includes a scenario modeling unit, a scheme simulation unit, a performance analysis unit, and a parameter optimization unit; before the task is issued, it simulates the operation based on the dynamic capability model and the task decomposition results, with a simulation prediction error of ≤3%, and can output performance indicators such as task completion time, resource consumption rate, and failure risk rate.

[0015] Preferably, the dynamic weight adjustment rule of the intelligent task allocation module is as follows: when the task phase requires high movement speed, adjust... =0.55、 =0.35、 =0.1; When the task phase requires high operational precision, adjust... =0.35、 =0.35、 =0.3; dynamically adapt to changes in task requirements through weighting.

[0016] This invention provides a method and system for heterogeneous robot collaboration. It has the following beneficial effects: 1. This invention achieves low-latency data transmission and significantly enhances link stability through a multi-protocol adaptive communication adaptation layer and an improved latency compensation model. This results in a breakthrough improvement in communication adaptation performance, supports seamless interaction of multiple mainstream protocols, and effectively solves the problems of poor protocol compatibility and insufficient communication reliability.

[0017] 2. This invention significantly optimizes the intelligence level of task allocation and the efficiency of resource utilization. Based on the improved algorithm and multi-objective optimization logic, combined with the dynamic capability perception mechanism, it achieves precise matching between tasks and robots, resulting in more balanced resource consumption and significantly improved task execution efficiency.

[0018] 3. The collaborative positioning accuracy of this invention reaches the advanced level in the industry. Relying on the multi-source fusion positioning and dynamic calibration mechanism, positioning drift is effectively suppressed, and positioning stability in complex environments is greatly improved, providing a solid guarantee for high-precision collaboration of heterogeneous robots.

[0019] 4. This invention adopts a multi-mode fault detection and intelligent dynamic reconstruction mechanism, which enables rapid fault identification, efficient and accurate task redistribution, and full guarantee of task execution continuity, thus significantly improving the reliability and anti-interference capability of the collaborative system. Attached Figure Description

[0020] Figure 1 This is a flowchart of the present invention; Figure 2 This is a system diagram of the present invention. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] Example: Please see the appendix Figure 1 - Appendix Figure 2 This invention provides a heterogeneous robot collaboration method and system, comprising the following steps: Step 1: Construct a multi-protocol adaptive communication adaptation layer. Based on an intelligent protocol conversion unit, realize heterogeneous robot data interaction using CAN, Ethernet / IP, ROS, and Modbus protocols. Control communication latency through an improved latency compensation model. ≤45ms; The calculation formula for the improved delay compensation model is as follows: ,in, For communication distance, For data bandwidth, This represents the protocol complexity coefficient; the parameter value is... =0.018ms / m, =0.0008ms / MB, =0.5ms; Transmission delay is offset by a two-stage algorithm of pre-compensation and real-time calibration; The communication adaptation module is activated, and the intelligent protocol parsing unit automatically identifies the communication protocol types of heterogeneous robots. This unit has a built-in protocol feature library that can quickly match the frame structure, data encoding rules, and transmission timing characteristics of four mainstream protocols: CAN, Ethernet / IP, ROS, and Modbus, achieving millisecond-level protocol type identification. Subsequently, the dynamic data format conversion unit converts the heterogeneous data from different protocols into a standardized data frame format, including data field alignment, byte order adjustment, and redundant information removal, ensuring the consistency of data interaction between heterogeneous robots. Step 2: Collect real-time status parameters and dynamic capability factors for each heterogeneous robot. Status parameters include remaining battery power S, current load L, movement speed v, and device health H. Dynamic capability factors include motion adaptation coefficient M and operation accuracy coefficient P. The weighted calculation formula for equipment health H is: ,in, =0.45, =0.25, =0.15, =0.15, and =1; For hardware health status, This refers to the software's running status. For environmental adaptation status, Task fatigue level; The status and capability perception module synchronously activates each detection unit to comprehensively collect the operating status and capability characteristics of each heterogeneous robot. The high-precision power detection unit uses the Coulomb counting method combined with open-circuit voltage calibration technology to collect the robot's remaining power S in real time, with a sampling frequency set to 1Hz and a measurement error controlled within ±1%. The load dynamic monitoring unit detects changes in the current and torque of the robot's actuators and calculates the current load L by combining it with a preset load-electrical signal mapping model. The speed acquisition unit integrates wheel encoder data and IMU speed information, and uses a Kalman filter algorithm for smoothing to obtain a high-precision motion speed v. Step 3: Based on the improved particle swarm optimization-auction fusion algorithm, construct the optimization objective function for the task allocation unit: in, , , ,and ; This represents the total time required to complete the task. The variance of resource consumption for each robot, This function determines the task-robot fit and enables dynamic task allocation. The key parameter settings for the improved particle swarm-auction fusion algorithm are as follows: Inertia weight calculation formula: ,in, =0.95, =0.35; This represents the current iteration number. This represents the maximum number of iterations. Learning factors: c1=2.1, c2=2.1; Auction bidding threshold: 0.75; Particle population size: 40-60; Algorithm convergence accuracy ≤5e-4; After receiving the total task, the intelligent task allocation module first decomposes it into multiple independent atomic tasks. Then, based on the improved particle swarm optimization-auction fusion algorithm, it assigns the optimal execution robot to each atomic task. The core of this algorithm is the multi-objective optimization objective function F=α・T+β・R+γ・C, where T is the total task completion time (calculated based on robot movement speed and task complexity), R is the variance of resource consumption of each robot (reflecting the balance of resource allocation), and C is the task-robot fit (calculated by the degree of matching between the mobility fit coefficient M, the operational accuracy coefficient P, and the task requirements), and α+β+γ=1, with dynamic weight coefficients adjusted according to the characteristics of the task stage. Step 4: Employ a vision-IMU-LiDAR-ultrasound multi-source fusion positioning method, combined with relative position constraints and dynamic calibration mechanisms among heterogeneous robots; Multi-source fusion localization employs an adaptive extended Kalman filter. The core equations and parameters are as follows: State equation: ; Observation equation: ,in, The position-attitude-velocity state vector, To control the input, For process noise, For observation noise; initial process noise covariance matrix: Q=diag([8e-5,8e-5,8e-5,5e-7,5e-7,5e-7]); initial observation noise covariance matrix: R=diag([8e-4,8e-4,8e-4]); Q and R are dynamically updated every 10 sampling periods to adapt to environmental changes; After the high-precision collaborative positioning module is activated, the multi-source data acquisition unit synchronously acquires raw data from the visual sensor (image feature points), IMU (angular velocity, acceleration), lidar (distance, angle) and ultrasonic sensor (near-range ranging). The time stamp synchronization technology is used to achieve spatiotemporal alignment of the multi-source data, with a synchronization error of ≤1ms. Step 5: Monitor the robot's operating status in real time. When a fault occurs, trigger the intelligent dynamic reconfiguration unit to complete the task reassignment of the faulty robot within 80ms. The intelligent fault reconfiguration module monitors the robot's operating status in real time through a dual-mode fault detection unit, employing a multi-mode detection mechanism combining threshold method, trend analysis method, and abrupt change detection method. The sampling frequency is set to 15Hz. Failure to detect a status parameter (such as battery level, load, speed, or health status) will result in the robot exceeding a preset threshold for four consecutive sampling cycles, or if the absolute value of the parameter's trend slope is ≥0.7, or if the parameter's abrupt change amplitude is ≥20%. When this occurs, it is determined to be a fault.

[0023] A heterogeneous robot collaborative system, comprising: The communication adaptation module includes an intelligent protocol parsing unit, a dynamic data format conversion unit, and an improved delay compensation unit, controlling the communication delay τ≤45ms. The module adopts an FPGA+MCU hardware architecture, with the FPGA responsible for high-speed protocol parsing and data format conversion, and the MCU responsible for delay compensation algorithm execution and module coordination. The status and capability perception module integrates multiple high-precision sensors and communicates with the robot's main control unit via a CAN bus to achieve centralized data acquisition and preprocessing. The intelligent task allocation module is deployed on an edge computing node, using an ARM Cortex-A9 processor to ensure efficient algorithm operation. The high-precision collaborative positioning module transmits multi-source sensor data at high speed via a PCIe interface, and the fusion filtering algorithm is hardware-accelerated in the FPGA. The intelligent fault reconstruction module communicates with the central control module via Ethernet. The central control module uses an industrial-grade embedded computer to coordinate the work of all modules. Status and Capability Perception Module: Includes a high-precision power detection unit, a load dynamic monitoring unit, a speed acquisition unit, a multi-dimensional health assessment unit, and a capability factor calculation unit, which collects status parameters and dynamic capability factors; Intelligent task allocation module: includes algorithm operation unit, parameter input unit, allocation execution unit, and bid evaluation unit, based on improved particle swarm optimization-auction fusion algorithm and objective function. Achieve dynamic allocation; High-precision collaborative positioning module: includes multi-source data acquisition unit, adaptive fusion filtering unit, dynamic positioning calibration unit, and relative position constraint unit, achieving positioning accuracy ≤1.5cm; Intelligent fault reconstruction module: includes dual-mode fault detection unit, rapid task redistribution unit, status feedback unit, and task continuity guarantee unit, with a fault response time ≤80ms; Central control module: includes instruction generation unit, data storage unit, logic judgment unit, and global optimization scheduling unit, and communicates bidirectionally with the above modules and coordinates them in an overall manner; The adaptive communication topology management unit includes a neighbor discovery unit, a link quality assessment unit, a route optimization unit, and a topology reconstruction unit. It constructs a dynamic mesh network topology. The neighbor discovery module of the adaptive communication topology management unit uses a periodic broadcast probe mechanism, sending a probe packet every 200ms to discover other robot nodes within the communication range. The link quality assessment unit calculates... Assess link stability when When the value is less than 0.6, the routing optimization unit triggers the topology reconstruction algorithm, which uses the distributed minimum spanning tree algorithm to reconstruct the communication topology to ensure link reliability. Before the task is issued, the simulation verification and performance evaluation module simulates the robot collaboration process based on the virtual scenario constructed by the scenario modeling unit. The performance analysis unit outputs indicators such as task completion time, resource consumption rate, and failure risk rate to provide a basis for optimizing the actual collaboration scheme.

[0024] The judgment rule of the dual-mode fault detection unit is as follows: a fault is determined when a certain state parameter meets any of the following conditions: the monitoring threshold of claim 3 is exceeded for 4 consecutive sampling periods; the absolute value of the slope of the state parameter change trend is ≥0.7; the amplitude of the sudden change of the state parameter is ≥20%.

[0025] The link quality assessment unit calculates the link quality indicator using the following formula: ,in, For receiving signal strength indication, For bit error rate, Transmission delay; weighting factor values =0.4, =0.35, γ=0.25; when The topology reconstruction algorithm is triggered when the value is less than 0.6.

[0026] It also includes a simulation verification and performance evaluation module, which includes a scenario modeling unit, a scheme simulation unit, a performance analysis unit, and a parameter optimization unit. Before the task is issued, it simulates the operation based on the dynamic capability model and the task decomposition results. The simulation prediction error is ≤3%, and it can output performance indicators such as task completion time, resource consumption rate, and failure risk rate.

[0027] The dynamic weight adjustment rule of the intelligent task allocation module is as follows: when the task phase requires high movement speed, adjust... =0.55、 =0.35、 =0.1; When the task phase requires high operational precision, adjust... =0.35、 =0.35、 =0.3; dynamically adapt to changes in task requirements through weighting.

[0028] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A heterogeneous robot collaboration method, characterized by, Comprising the following steps: Step one: Constructing multi-protocol adaptive communication adaptation layer, based on intelligent protocol conversion unit to realize heterogeneous robot data interaction of CAN, Ethernet / IP, ROS and Modbus protocol, and control communication delay through improved delay compensation model ≤ 45 ms; Step two: Collect real-time state parameters and dynamic capability factors of each heterogeneous robot, the state parameters include remaining power S, current load L, motion speed v, device health degree H, the dynamic capability factors include movement adaptation coefficient M, operation precision coefficient P; Step three: the task allocation unit based on the improved particle swarm-auction fusion algorithm, the optimization objective function is constructed: Wherein, , , , and ; is the total task completion time, is the variance of each robot resource consumption, is the task-robot fitness; through the function, dynamic task allocation is realized; Step four: Use a multi-source fusion positioning method of vision-IMU-laser radar-ultrasonic wave, combined with relative position constraints and dynamic calibration mechanism between heterogeneous robots; Step five: Real-time monitoring of robot operating state, triggering intelligent dynamic reconstruction unit when fault occurs, completing task redistribution of fault robot within 80ms.

2. The heterogeneous robot collaboration method of claim 1, wherein, The calculation formula of the improved delay compensation model is: , wherein, is a communication distance, is a data bandwidth, is a protocol complexity coefficient; the parameter values are = 0.018 ms / m, = 0.0008 ms / MB, = 0.5 ms; the transmission delay is offset by a pre-compensation-real-time calibration two-stage algorithm.

3. The heterogeneous robot collaboration method of claim 1, wherein, The weighted calculation formula of the device health degree H is: , wherein, = 0.45, = 0.25, = 0.15, = 0.15, and = 1; is a hardware health state, is a software running state, is an environment adaptation state, is a task execution fatigue degree.

4. The heterogeneous robot collaboration method of claim 1, wherein, The key parameter settings of the improved particle swarm-auction fusion algorithm are as follows: an inertia weight calculation formula: , wherein, = 0.95, = 0.35; is a current iteration number, is a maximum iteration number; Learning factors: c1=2.1, c2=2.1; Auction bid threshold: 0.75; Particle population size: 40-60, algorithm convergence accuracy ≤5e-4.

5. The heterogeneous robot collaboration method of claim 1, wherein, In step four, the multi-source fusion positioning adopts adaptive extended Kalman filter, and the core equation and parameters are as follows: state equation: observation equation: , wherein, is a position-pose-velocity state vector, is a control input, is a process noise, is an observation noise; an initial process noise covariance matrix: Q=diag([8e-5,8e-5,8e-5,5e-7,5e-7,5e-7]); an initial observation noise covariance matrix: R=diag([8e-4,8e-4,8e-4]; dynamically update Q and R every 10 sampling periods to adapt to environmental changes.

6. A heterogeneous robot collaboration system using the heterogeneous robot collaboration method according to any one of claims 1 to 5, characterized by, Comprising: Communication adaptation module: containing intelligent protocol analysis unit, dynamic data format conversion unit, improved delay compensation unit, controlling communication delay τ≤45ms; State and capability perception module: containing high-precision power detection unit, load dynamic monitoring unit, speed acquisition unit, multi-dimensional health assessment unit, capability factor calculation unit, collecting state parameters and dynamic capability factors; The intelligent task allocation module comprises an algorithm operation unit, a parameter input unit, an allocation execution unit and a bid evaluation unit, and is based on an improved particle swarm-auction fusion algorithm and a target function Dynamic allocation is realized High-precision cooperative positioning module: containing multi-source data acquisition unit, adaptive fusion filtering unit, dynamic positioning calibration unit, relative position constraint unit, achieving positioning accuracy ≤1.5cm; Intelligent fault reconstruction module: containing dual-mode fault detection unit, fast task redistribution unit, state feedback unit, task continuity guarantee unit, fault response time ≤80ms; Central control module: containing instruction generation unit, data storage unit, logic judgment unit, global optimization scheduling unit, bidirectional communication with the above modules and overall coordination; Adaptive communication topology management unit: containing neighbor discovery unit, link quality assessment unit, routing optimization unit, topology reconstruction unit, constructing dynamic Mesh network topology.

7. A heterogeneous robotic collaboration system according to claim 6, wherein, The determination rule of the dual-mode fault detection unit is: when any of the following conditions is met, it is determined to be a fault: exceeding the monitoring threshold of claim 3 for 4 consecutive sampling periods; the absolute value of the state parameter change trend slope is ≥0.7; the state parameter mutation amplitude is ≥20%.

8. A heterogeneous robotic collaboration system according to claim 6, wherein, The link quality assessment unit calculates the link quality indicator by the following formula: wherein is the received signal strength indicator, is the bit error rate, is the transmission delay; the weighting factors take the values = 0.4, = 0.35, γ = 0.25; the topology reconfiguration algorithm is triggered when < 0.

6.

9. The heterogeneous robotic collaboration system of claim 6, wherein, It also includes simulation verification and performance evaluation module, containing scene modeling unit, scheme simulation unit, performance analysis unit, parameter optimization unit; based on dynamic capability model and task decomposition result before task issuing, simulating operation, simulation prediction error ≤3%, can output performance indicators such as task completion time, resource consumption rate, fault risk rate.

10. The heterogeneous robotic collaboration system of claim 6, wherein, The dynamic weight adjustment rule of the intelligent task allocation module is: when the task stage requires high moving speed, adjusting = 0.55, = 0.35, = 0.1; when the task stage requires high operation precision, adjusting = 0.35, = 0.35, = 0.3; the weight is dynamically adapted to the change of task demand.

Images