Modular robot and control system thereof

Abstract

The invention relates to the technical field of robots, in particular to a modular robot and a control system thereof. According to the technical scheme, the modular robot is provided with four functional modules including a power module, a sensing module, an execution module and an energy module, and a control system of the modular robot comprises a task composer, a self-adaptive feedback loop module and an open communication interface. And task-level structure scheduling, state adaptive control and real-time information exchange between modules and between robots are realized respectively. The modular robot system provided by the invention has a multi-modal movement capability, supports task-level adaptive arrangement, has a feedback learning capability, and can realize module-level intelligent collaboration under an open communication architecture.

Description

Technical Field

[0001] This invention relates to the field of technology, and more particularly to modular robots and their control systems. Background Technology

[0002] With the rapid development of artificial intelligence, automatic control, and modular manufacturing technologies, modular robot systems have gradually become a research hotspot in various fields such as intelligent manufacturing, disaster relief, underground mining, and complex environment exploration due to their high flexibility, scalability, and rapid reconfiguration capabilities. Most existing modular robots employ fixed motion patterns and centralized scheduling strategies, making it difficult to simultaneously handle multi-task concurrent scheduling and adapt to complex terrain. Furthermore, they still face the following technical bottlenecks in areas such as group collaboration, motion control, and task switching: Limited movement patterns and poor terrain adaptability: Most modular robots only have a single motion mechanism (such as tracks or differential wheels), resulting in low efficiency and poor reliability when moving between different terrains, and lack of multimodal motion capabilities; Weak closed-loop feedback capability of control systems: Existing control systems generally rely on static parameter models, which are difficult to adapt to dynamic changes in tasks, lack real-time adaptive control mechanisms based on sensor information, and have insufficient system robustness. Weak task scheduling and module decoupling: Traditional task planning systems have failed to be deeply integrated with the modular structure of robots. The task orchestration mechanism relies on preset processes and cannot dynamically adjust the task flow graph structure according to the module status. Closed information communication protocols limit group collaboration capabilities: Closed communication protocols are typically used between modules and between multiple robots, making it difficult to achieve interconnection of heterogeneous modules, group-level task collaboration, and remote strategy updates.

[0003] Therefore, we propose a modular robot and its control system to address the existing problems. Summary of the Invention

[0004] The purpose of this invention is to address the problems existing in the background art by proposing a modular robot and its control system.

[0005] To achieve the above object, the present invention provides the following technical solutions: Modular robots, including: Each of the functional modules has an independent structural housing and electrical interface, and is used to realize a preset function. The functional modules include a power module, a sensing module, an execution module and an energy module. A set of standardized mechanical connection devices is disposed at the end of the functional module to enable rapid assembly and decoupling between any two functional modules. The standardized mechanical connection devices include a self-aligning positioning structure and a magnetic or mechanical buckle mechanism with locking function, allowing module assembly and disassembly without tools. A set of self-identification interfaces is used for automatic identification and topology awareness between modules, including identification of module number, connection direction and function type; A bus communication channel, which connects all the modules, is used for the unified transmission of control commands, power and data signals; One of the central management modules has a built-in processor for dynamically generating control logic based on the combined structure of the connected modules and coordinating the collaborative work of the modules.

[0006] Preferably, the power module includes a motor unit that can output adjustable torque and integrates a position encoder and a speed sensor for real-time acquisition of the rotational state of the drive shaft, thereby achieving closed-loop control of the precise position and attitude of the execution module. The modular robot and its control system for the perception module include a lidar, an infrared ranging sensor, an inertial measurement unit (IMU), and a stereo vision camera. The sensing units work together to construct a three-dimensional map of the surrounding environment through a fusion algorithm. The energy module includes a rechargeable battery pack, an energy management unit, and a wireless energy receiving module. The energy management unit is used to automatically adjust the power output priority according to the real-time power consumption of each module and dynamically allocate power supply through the bus. The wireless energy receiving module supports the Qi standard contactless charging interface, enabling continuous energy replenishment without power interruption. The central management module is equipped with a fault-tolerant logic analysis unit, which automatically triggers a fault isolation mechanism when any functional module is detected to have an abnormal response, communication interruption, or performance degradation. This mechanism shields the output signal of the faulty module and reconstructs the control path to maintain the stable operation of the overall robot system.

[0007] Preferably, the standardized mechanical connection device includes a pair of foolproof plug-in structures and a locking device. The foolproof structure adopts a combination design of irregular positioning pins and limiting grooves, which only allows docking in a predefined direction. The locking device is a mechanical spring buckle, which supports automatic locking after module plugging and has an electronically controlled quick release function, improving loading and unloading efficiency and reliability.

[0008] The control system for the modular robot includes: One of the module identification units is used to generate a complete module topology diagram by reading the self-identification interface information of each functional module; One of the aforementioned structure perception and control mapping engines automatically matches a predefined control model or generates a new control strategy based on the module topology graph; One of the task orchestrators is used to receive external task instructions, parse them into subtasks, and assign them to corresponding functional modules; One of the aforementioned adaptive feedback loop modules is used to collect the operating status, fault information and execution deviation of each module in real time, and dynamically adjust control parameters or reconfigure task paths. One of the aforementioned open communication interfaces is used to connect with an external operating platform or cloud system to achieve remote monitoring, task updates, or collaborative control; The control system has the ability to identify and update modules for hot-swapping, and can maintain the overall task execution without interruption during the addition or removal of modules, thus maintaining system stability and continuity.

[0009] Preferably, the structure perception and control mapping engine has a built-in deep learning model that can automatically optimize the collaborative action logic and parameter settings between modules by analyzing historical task execution data, so that the control strategy has good transferability and adaptability when facing different tasks.

[0010] Preferably, the task orchestrator provides a graphical user interface that supports adding modules by icons, constructing task flowcharts by dragging and dropping lines, setting node parameters and defining conditional jumps. The generated task flow can be automatically translated into executable code, significantly reducing the operating threshold for non-professionals.

[0011] Preferably, the adaptive feedback loop module combines fuzzy logic control algorithm with neural network prediction mechanism to adjust the control strategy in real time according to environmental disturbances, sensor noise and execution result errors during system operation, effectively improving system response speed and dynamic robustness.

[0012] Preferably, the open communication interface conforms to the ROS (Robot Operating System) communication architecture standard, supports TCPROS and UDPROS communication protocols, and can realize high-speed data exchange and distributed control integration with other heterogeneous systems or third-party devices.

[0013] Preferably, the control system has a pre-set module function redundancy mapping table. After a module failure is detected, the system can reallocate task functions within the range of replaceable modules and automatically adjust the control path and execution logic, thereby ensuring that the robot can still maintain basic task operation under some fault conditions.

[0014] Preferably, the system can simultaneously identify the structure and status information of modular robot units, realize cluster scheduling and resource allocation through the robot coordination module, and support distributed task publishing, mutual navigation and faulty robot replacement functions, making it suitable for collaborative operation scenarios.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: Multimodal three-degree-of-freedom mobile platform design improves terrain adaptability and passage efficiency: This system supports rapid switching between various mobile mechanisms such as differential, Mecanum, and tracked. Combined with three-degree-of-freedom motion modeling, it can maintain stable movement and attitude control in complex terrains such as flat ground, obstacles, and slopes. By adopting multimodal state transition functions and control matching mechanisms, it can realize three-dimensional mapping of "task-terrain-motion mode". A task orchestrator is introduced to realize a scheduling mechanism of bidirectional matching between tasks and structures: Based on module topology recognition and capability modeling, the task orchestrator can perform functional mapping and task allocation for each module, realize task-level structure adaptation, and the orchestrator uses a directed acyclic graph (DAG) to describe the task dependency structure, which can dynamically adjust task priority and execution order. An adaptive feedback loop module is constructed to improve control accuracy and system robustness: The system integrates Kalman filtering, fuzzy control, and self-learning parameter adjustment mechanism to construct a full-process closed-loop control link; through real-time error estimation, control parameters and path planning are dynamically corrected to adapt to non-ideal factors such as load disturbances and environmental changes. An open communication interface architecture is adopted to improve module reusability and system scalability: the system communication interface is based on standardized protocols (such as CAN, EtherCAT, ROS2 DDS) to achieve cross-module / cross-platform data synchronization and control integration; it supports remote task issuance, policy updates, status feedback and collaborative control of group robots. Supports parallel operation of multiple robots in groups and hierarchical scheduling of tasks: The control system has the ability to fuse group-level states and distribute behaviors, and can realize the concurrent operation and dynamic adjustment of multiple robot tasks in scenarios such as disaster relief, multi-point sampling, and construction of cellular-like networks. Attached Figure Description

[0016] Figure 1 This is a logical sequence diagram of the present invention. Detailed Implementation

[0017] 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.

[0018] The modular robot proposed in this invention includes: Functional modules, each with an independent structural housing and electrical interface, are used to implement a preset function. The functional modules include a power module: the power module includes a motor unit that can output adjustable torque and integrates a position encoder and speed sensor to collect the rotation state of the drive shaft in real time, so as to realize closed-loop control of the precise position and attitude of the execution module, which is suitable for industrial or service scenarios that require high-precision motion response. Perception Module: The perception module includes at least one lidar, one infrared ranging sensor, one inertial measurement unit (IMU) and one stereo vision camera. The sensing units work together through a fusion algorithm to construct a three-dimensional map of the surrounding environment, which is used to realize multiple perception functions such as path planning, obstacle avoidance, autonomous positioning and target recognition. Energy Module: The energy module includes a rechargeable battery pack, an energy management unit, and a wireless energy receiving module. The energy management unit is used to automatically adjust the power output priority according to the real-time power consumption of each module and dynamically allocate power supply through the bus. The wireless energy receiving module supports the Qi standard contactless charging interface, enabling continuous power replenishment and module execution without power interruption.

[0019] A set of standardized mechanical connection devices, located at the ends of functional modules, enables rapid assembly and decoupling between any two functional modules. These standardized mechanical connection devices include a self-aligning positioning structure and a magnetic or mechanical latching mechanism with locking function, allowing module assembly and disassembly without tools. Each standardized mechanical connection device includes a pair of foolproof plug-in structures and a locking device. The foolproof structure uses a combination of a shaped positioning pin and a limiting groove, allowing only predefined orientations for docking. The locking device is an electrically controlled magnetic lock or a mechanical spring latch, supporting automatic locking after module insertion and featuring an electrically controlled rapid release function, improving loading and unloading efficiency and reliability.

[0020] A set of self-identification interfaces are used for automatic identification and topology awareness between modules, including identification of module number, connection direction and function type; A bus communication channel runs through all connected modules and is used to uniformly transmit control commands, power and data signals; A central management module with a built-in processor is used to dynamically generate control logic based on the combined structure of the connected modules and coordinate the collaborative work of each module. The central management module is equipped with a fault-tolerant logic analysis unit, which is used to automatically trigger a fault isolation mechanism when any functional module is detected to have an abnormal response, communication interruption or performance degradation, shield the output signal of the faulty module, and reconstruct the control path to maintain the stable operation of the overall robot system.

[0021] Example 1 like Figure 1 As shown, the modular robot control system proposed in this invention includes the following steps: Step 1: Module topology identification and structure encoding After the modules are connected, the control system generates a structural topology diagram by identifying the unique codes and connection relationships of each module. The coding vector of the topology diagram is represented as follows:

[0022] in: : No. The function number of each module; : Adjacency matrix between modules; : Structural topology vector, used for subsequent control mapping generation.

[0023] Related to subsequent steps: Structure encoding It will serve as the core input basis for subsequent motion models, perception configurations, control mappings, etc.

[0024] Step 2: Motion Model Establishment The structure perception and control mapping engine automatically matches predefined control models or generates new control strategies based on the module topology graph. The structure perception and control mapping engine has a built-in deep learning model that can automatically optimize the collaborative action logic and parameter settings between modules by analyzing multiple historical task execution data, so that the control strategy has good transferability and adaptability when facing different tasks.

[0025] The robot body adopts a multimodal mobile platform, considering a three-degree-of-freedom mobile scenario. Based on the topology, effective power module groups are extracted to generate a motion capability model of the overall platform. Its state equation modeling is as follows:

[0026] in: , , The robot's current position and orientation; Linear velocity and angular velocity input; Angular velocity about the vertical axis, i.e., the rate of change of the heading angle; : Sampling time interval.

[0027] Related to step 1: The dynamic model is automatically selected and modeled based on the structural topology.

[0028] Related to steps 4 and 5: Provides a dynamic basis for path planning and control feedback.

[0029] Step 3: Multi-sensor fusion localization algorithm The extended Kalman filter (EKF) algorithm is used to fuse IMU, visual, and laser data. Based on the type of perception module for structural recognition, the control system dynamically configures the perception fusion module.

[0030]

[0031] Parameter meaning: : Current optimal state estimate; Kalman gain; : Observation vector; : Jacobian matrix of the observation model; : Measure noise covariance; : Predicted value of state covariance.

[0032] Related to step 1: The sensing module is automatically activated; Related to step 2: The positioning results are fed back to the motion model for correction.

[0033] Step 4: Adaptive adjustment of controller parameters To improve control robustness, a fuzzy-PID hybrid control method is adopted, which automatically adjusts the control gain according to the error:

[0034] Among them, three gains , , All are outputs of fuzzy inference functions, based on error. With error change rate Make adjustments in real time.

[0035] Related to step 3: State estimation is used for error calculation; Related to step 5: Feedback is applied to path execution error correction.

[0036] Step 5: Path planning optimization Based on structural capabilities and environmental perception information, an improved A* algorithm is used to complete path planning. Its objective function is:

[0037] in: The cost from the starting point to the current node; : Estimate the distance from the current node to the target (heuristic value); Path complexity (e.g., turning frequency, local obstacle density); Complexity penalty factor, used to balance path smoothness and shortest path.

[0038] Related to steps 2, 3, and 4: Control capabilities, environmental data, and execution strategies all participate in path generation and adjustment.

[0039] Step 6: Redundant Module Scheduling Strategy If any functional module (such as sensing or actuation) fails during task execution, the control system can quickly select a replacement module using a scoring function.

[0040] in: Functional compatibility; Performance redundancy; : Cost of task migration; , , Adjustable weighting coefficients, satisfying + + =1.

[0041] Related to step 1: Predefined redundant relationships in structure identification; Linked with steps 2–5: Remodel the control process after system reconstruction.

[0042] Step 7: Cluster control and collaborative task allocation For a cluster of multiple modular robots, task allocation is optimized using minimum total cost matching:

[0043] in: : All possible matches between the task and the robot; :robot Complete the task The cost (calculated based on a weighted average of distance, load capacity, and task complexity).

[0044] Linked with step 6: Task allocation takes into account the currently available robots and their capability status; Related to step 8: After each task is assigned, proceed to the task scheduling flow graph.

[0045] Step 8: Task flow graph modeling and scheduling The task flow is modeled using a directed acyclic graph (DAG), with each task node... The priority is calculated by the following function:

[0046] in: Task weight; The current task queue is delayed; Module execution capability coefficient; Small constants that should be avoided from being divided by zero.

[0047] Related to steps 5 and 7: Specific task execution scheduling after the path is completed; Related to step 9: Task execution feedback is used for reinforcement learning optimization.

[0048] Step 9: Self-learning control strategy update The system updates the control policy based on the execution trajectory after each task execution using a policy gradient algorithm (such as DDPG).

[0049] in: : Current policy network parameters; Execution trajectory; Instant rewards for every step; Discount factor: controls the degree of focus on long-term returns.

[0050] In conjunction with steps 4, 5, and 8: the control and scheduling strategies are continuously iterated and optimized.

[0051] Step 10: Inter-module communication delay compensation model In high-frequency communication, to reduce state drift caused by latency, a discretized Smith predictor model is adopted:

[0052] in: : The compensated predicted output value; Discrete transfer function of the system; Communication delay steps; Current control input.

[0053] Linked with step 7: Improve the synchronization accuracy of multi-machine scheduling and control.

[0054] The task orchestrator receives external task instructions, parses them into subtasks, and assigns them to corresponding functional modules. The task orchestrator provides a graphical user interface that supports adding modules by icons, building task flowcharts by dragging and dropping lines, and supports setting node parameters and defining conditional jumps. The generated task flow can be automatically translated into executable code, significantly reducing the operating threshold for non-professionals.

[0055] The task orchestrator is the core scheduling module of the control system. It is responsible for task decomposition, priority sorting, process scheduling and module invocation, and coordinates multiple robots and their internal modules to work together to complete complex tasks.

[0056] After parsing the module topology in step 1, the task orchestrator generates a capability table based on the recognition results to determine the types of tasks that the module can undertake. In step 2, a motion model is established based on the motion capability, and an adaptive path generator or planning algorithm is assigned. In step 4, the control module (such as a fuzzy controller) is called and initialized to ensure a closed loop in the control process; In step 5, the path calculation priority is arranged and its execution is controlled to be synchronized with the task scheduling flow graph; In step 7, parallel tasks are issued to multiple robot instances, and the task progress status is tracked. In step 8, it is directly responsible for task DAG generation and scheduling control; it also works in conjunction with the feedback module to handle dynamic priority adjustments.

[0057] The adaptive feedback loop module is used to collect the operating status, fault information and execution deviation of each module in real time, and dynamically adjust the control parameters or reconfigure the task path. The adaptive feedback loop module adopts a combination of fuzzy logic control algorithm and neural network prediction mechanism. During the system operation, it adjusts the control strategy in real time according to environmental disturbances, sensor noise and execution result errors, effectively improving the system response speed and dynamic robustness.

[0058] The adaptive feedback loop module realizes a self-feedback process of closed-loop sensing → state estimation → control strategy update → behavior execution. It can dynamically adjust control parameters, path planning or task priority based on sensor data or task execution results.

[0059] In step 3, the adaptive feedback loop module obtains input from the sensing module and forms a state estimate through methods such as EKF, thus forming the first loop of the feedback loop. In step 4, state error information is received and PID parameters are dynamically adjusted to achieve closed-loop control behavior; When the detection module's function deteriorates or malfunctions in step 6, the feedback loop triggers a redundant module switching command; In step 9, the error data accumulated in the loop is converted into reward signals in reinforcement learning to optimize subsequent policies; In step 10, the communication delay behavior is modeled and fed back into the predictive controller to compensate for the behavior delay error.

[0060] The adaptive feedback loop module spans the entire process of perception, control, and execution, and is a key mechanism to ensure system stability and adaptability.

[0061] An open communication interface is used to connect with external operating platforms or cloud systems to achieve remote monitoring, task updates or collaborative control. The open communication interface conforms to the ROS (Robot Operating System) communication architecture standard and supports the TCPROS and UDPROS communication protocols, enabling high-speed data exchange and distributed control integration with other heterogeneous systems or third-party devices.

[0062] The open communication interface supports heterogeneous communication protocol interaction between modules, robots, and external systems, ensuring the real-time performance and universality of data sharing, task collaboration, and control command transmission.

[0063] In step 1, the open communication interface reads the structure, ID, and status data from each module through the interface to achieve identification and initialization. In step 3, the sensed data is transmitted to the control unit via a communication bus for fusion. In step 7, the robots in the group exchange status and task data through a communication interface, supporting distributed collaboration; In step 9, the control strategy model can be remotely uploaded / downloaded, supporting OTA self-learning and model migration; In step 10, the real-time communication protocol provides message timestamps and network prediction models, which are then input into the latency modeler.

[0064] The control system has the capability of hot-swappable module identification and control update, which can maintain the overall task execution without interruption during the addition or removal of modules, thus maintaining system stability and continuity. The control system has a preset module function redundancy mapping table. After detecting the failure of a module, it can redistribute task functions within the range of replaceable modules and automatically adjust the control path and execution logic, thereby ensuring that the robot can still maintain basic task operation under partial failure conditions. The system can simultaneously identify the structure and status information of multiple modular robot units, realize cluster scheduling and resource allocation through multi-robot coordination modules, and support distributed task release, mutual navigation and faulty robot replacement functions, making it suitable for collaborative operation scenarios.

[0065] The above specific embodiments are merely several preferred embodiments of the present invention. Based on the technical solutions of the present invention and the relevant teachings of the above embodiments, those skilled in the art can make various alternative improvements and combinations to the above specific embodiments.

[0066] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, it is intended that all variations falling within the meaning and scope of equivalents of the claims be included within the present invention.

Claims

1. A modular robot, characterized in that, include: Each of the functional modules has an independent structural housing and electrical interface, and is used to realize a preset function. The functional modules include a power module, a sensing module, an execution module and an energy module. A set of standardized mechanical connection devices is disposed at the end of the functional module to enable rapid assembly and decoupling between any two functional modules. The standardized mechanical connection devices include a self-aligning positioning structure and a magnetic or mechanical buckle mechanism with locking function, allowing module assembly and disassembly without tools. A set of self-identification interfaces is used for automatic identification and topology awareness between modules, including identification of module number, connection direction and function type; A bus communication channel, which connects all the modules, is used for the unified transmission of control commands, power and data signals; One of the central management modules has a built-in processor for dynamically generating control logic based on the combined structure of the connected modules and coordinating the collaborative work of the modules.

2. The modular robot according to claim 1, characterized in that: The power module includes a motor unit that can output adjustable torque and integrates a position encoder and a speed sensor to collect the rotational state of the drive shaft in real time, thereby achieving closed-loop control of the precise position and attitude of the execution module. The modular robot and its control system for the perception module include a lidar, an infrared ranging sensor, an inertial measurement unit (IMU), and a stereo vision camera. The sensing units work together to construct a three-dimensional map of the surrounding environment through a fusion algorithm. The energy module includes a rechargeable battery pack, an energy management unit, and a wireless energy receiving module. The energy management unit is used to automatically adjust the power output priority according to the real-time power consumption of each module and dynamically allocate power supply through the bus. The wireless energy receiving module supports the Qi standard contactless charging interface, enabling continuous energy replenishment without power interruption. The central management module is equipped with a fault-tolerant logic analysis unit, which automatically triggers a fault isolation mechanism when any functional module is detected to have an abnormal response, communication interruption, or performance degradation. This mechanism shields the output signal of the faulty module and reconstructs the control path to maintain the stable operation of the overall robot system.

3. The modular robot according to claim 1, characterized in that: The standardized mechanical connection device includes a pair of foolproof plug-in structures and a locking device. The foolproof structure adopts a combination design of irregular positioning pins and limiting grooves, which only allows docking in a predefined direction. The locking device is a mechanical spring buckle, which supports automatic locking after module plugging and has an electronically controlled quick release function, improving loading and unloading efficiency and reliability.

4. A control system for the modular robot of claim 1, characterized in that, include: One of the module identification units is used to generate a complete module topology diagram by reading the self-identification interface information of each functional module; One of the aforementioned structure perception and control mapping engines automatically matches a predefined control model or generates a new control strategy based on the module topology graph; One of the task orchestrators is used to receive external task instructions, parse them into subtasks, and assign them to corresponding functional modules; One of the aforementioned adaptive feedback loop modules is used to collect the operating status, fault information and execution deviation of each module in real time, and dynamically adjust control parameters or reconfigure task paths. One of the aforementioned open communication interfaces is used to connect with an external operating platform or cloud system to achieve remote monitoring, task updates, or collaborative control; The control system has the ability to identify and update modules for hot-swapping, and can maintain the overall task execution without interruption during the addition or removal of modules, thus maintaining system stability and continuity.

5. The control system for the modular robot according to claim 4, characterized in that: The structure perception and control mapping engine has a built-in deep learning model that can automatically optimize the collaborative action logic and parameter settings between modules by analyzing historical task execution data.

6. The control system of the modular robot according to claim 4, characterized in that: The task orchestrator provides a graphical user interface that supports adding modules via icons, constructing task flowcharts by dragging and dropping lines, setting node parameters and defining conditional jumps, and automatically translating the generated task flow into executable code.

7. The control system for the modular robot according to claim 4, characterized in that: The adaptive feedback loop module combines fuzzy logic control algorithm with neural network prediction mechanism. During system operation, it adjusts the control strategy in real time according to environmental disturbances, sensor noise and execution result errors, effectively improving the system response speed and dynamic robustness.

8. The control system of the modular robot according to claim 4, characterized in that: The open communication interface conforms to the ROS (Robot Operating System) communication architecture standard, supports TCPROS and UDPROS communication protocols, and can realize high-speed data exchange and distributed control integration with other heterogeneous systems or third-party devices.

9. The control system for the modular robot according to claim 4, characterized in that: The control system has a pre-set module function redundancy mapping table. When a module fails, it can reallocate task functions within the range of replaceable modules and automatically adjust the control path and execution logic, thereby ensuring that the robot can still maintain basic task operation under some fault conditions.

10. The control system of the modular robot according to claim 4, characterized in that: The system can simultaneously identify the structure and status information of modular robot units, realize cluster scheduling and resource allocation through the robot coordination module, and support distributed task publishing, mutual navigation and faulty robot replacement functions, making it suitable for collaborative operation scenarios.

Images