Modular robot assembling and repairing system based on cooperation of multiple mechanical arms

Abstract

The invention discloses a modular robot assembling and repairing system based on cooperation of multiple mechanical arms, and the system comprises a modular robot subsystem which integrates a fault sensing module, a butt joint and communication module, a task execution module and a charging and power supply module, and each modular robot is equipped with a starting self-inspection mechanism and supports various peripheral extensions; a robot assembly is formed through butt joint of magnetic attraction type spring needle connectors; the multi-mechanical-arm cooperation subsystem adopts an offline-online double-layer planning framework and introduces a pairwise coordination and global coordination mechanism for a multi-mechanical-arm scene; and the central monitoring and diagnosis subsystem is used for carrying out environment modeling and mechanical arm planning calculation in real time and providing visual monitoring, service life prediction and fault diagnosis functions. The technical bottlenecks that a traditional modular robot system is insufficient in autonomy, single in function, weak in fault sensing capacity and long in fault recovery time are broken through, and the autonomous operation capacity of the robot in the extreme environment is effectively improved.

Description

Technical Field

[0001] This invention relates to the field of robotics, and in particular to a modular robot assembly and repair system based on multi-arm collaboration. Background Technology

[0002] In recent years, robotics technology has been widely applied in fields such as industrial automation, medical services, and smart logistics. However, traditional robots are typically designed specifically for the task requirements of particular fields; for example, welding robots are used for precision welding operations, while handling robots are suitable for logistics loading scenarios. However, as robot applications evolve towards greater complexity and dynamism, this task-specific robot design pattern has gradually revealed bottlenecks such as insufficient task adaptability and limited functional expansion. Whenever new application requirements arise, developers need to modify the robot's hardware and software architecture to some extent. This modification process not only increases development costs but also affects the deployment efficiency of robots in various scenarios. To address the problem of fixed functions in traditional robot systems, modular robots have emerged, enabling flexible expansion of robot functions through the reconfigurability of robot components.

[0003] In practical applications, the assembly of existing modular robots mainly relies on manual operation or single-arm systems. Manual assembly suffers from significant drawbacks such as low efficiency and high labor costs; while single-arm assembly, although achieving a degree of automation, has a relatively simple technical architecture and suffers from limitations such as limited workspace, inability to operate multiple modules simultaneously, and a lack of redundancy and fault tolerance mechanisms, making it inadequate for handling the complex independent assembly requirements of modular robots. In contrast, using multi-arm collaborative assembly of modular robots is a more effective option. However, combining multi-arm collaborative technology with modular robots still faces several challenges in practice.

[0004] On the one hand, existing modular robot designs have three shortcomings: First, they lack an "independent self-inspection mechanism for modules." Modular robots lack an independent self-inspection and evaluation mechanism before assembly, leading to potentially abnormal modular robots being directly installed into the robot assembly, only to be identified when the assembly malfunctions. This not only causes task interruption but may also trigger a series of chain reactions, causing more serious damage. Second, they lack peripheral interfaces or have low standardization of peripheral interfaces, limiting functional expandability. Existing technologies often use basic geometric shapes such as spheres or cubes as the uniform shape of modular robots and simplify or eliminate the peripheral interfaces found in conventional robots. While this design facilitates physical docking, it limits the robot's functional expandability, making it difficult to complete complex tasks through the combination of various peripheral components. For example, patent CN115366083A discloses a modular self-reconfigurable robot with a spherical body connected via magnetic attraction. While this design reduces the difficulty of physical docking between robots, it does not consider adding peripheral interfaces, thus limiting its perception, movement, operation, and interaction capabilities to some extent, making it difficult to meet the needs of complex and ever-changing application scenarios. Third, there is a lack of an "automated fault module replacement mechanism." For example, patent CN119556606A discloses a modular agricultural intelligent robot collaborative control system based on Beidou navigation and 5G. This patent effectively solves the problems of single-machine operation and low efficiency in traditional agricultural machinery through modular design. However, the modular components of this system can only be installed and disassembled manually. When a module fails, the system can only replace the faulty module manually, increasing time and labor costs.

[0005] On the other hand, multi-robotic arm collaborative operation technology also has obvious shortcomings in practical applications.

[0006] Currently, multi-arm collaborative operation modes are widely adopted in the field of industrial automation, and their collaborative strategies are mainly divided into two categories: centralized methods and decoupled methods. However, the computational complexity of centralized methods increases exponentially with the degrees of freedom and the number of robotic arms, making it difficult to balance computational speed and quality in dynamically changing real-world application scenarios. Decoupled methods are mostly only applicable to dual-arm collaborative scenarios and cannot be extended to multi-arm systems with three or more arms. If applied to multi-arm systems, the lack of a global coordination algorithm can easily lead to conflicts in local optima, resulting in motion deadlock. For example, patent CN119369420A discloses a method and system for collaborative motion planning of dual-arm robots for multi-arm cooperation. This patent solves the collision risk problem of traditional dual-arm robots in a shared workspace and the planning complexity problem brought about by high-dimensional joint configuration space by constructing a machine-machine-environment fusion collision information estimation model. However, the application scope of this patent is limited to collaborative motion planning of dual-arm robots, and there is no related technology such as a global planning algorithm to support its reliable extension to multi-arm collaborative scenarios with three or more arms.

[0007] In summary, existing technologies have shortcomings in modular robot self-inspection capabilities, peripheral interface standardization, automatic fault repair, and multi-arm collaborative operation. Therefore, there is an urgent need to provide a modular robot assembly and repair system based on multi-arm collaboration. Summary of the Invention

[0008] The purpose of this invention is to address the shortcomings of existing technologies by proposing a modular robot assembly and repair system based on multi-arm collaboration.

[0009] To achieve the above objectives, the present invention adopts the following technical solution:

[0010] A modular robot assembly and repair system based on multi-manipulator collaboration includes: a modular robot subsystem, which acts as the object of manipulator operation and docks to form a robot assembly; a multi-manipulator collaboration subsystem, which is used for robot assembly and repair operations; and a central monitoring and diagnostic subsystem, which is used for system status monitoring and diagnosis.

[0011] The present invention is further configured such that the modular robot subsystem includes: a task execution module for the robot to complete its assigned tasks; a fault perception module for sensing whether the robot has malfunctioned; a docking and communication module for realizing the docking function between communication and the modular robot; and a charging and power supply module for charging and powering the robot.

[0012] The task execution module includes optional motion execution components, vision components, motion components, perception components, and special-purpose components: the motion execution components include end effectors for performing various operational tasks; the vision components include vision acquisition and processing instruments for performing visual tasks; the motion components include a chassis system for performing various motion tasks; the perception components include sensors for assisting the robot in perceiving its own state and surrounding environment; and the special-purpose components include components designed for specific working environments or task requirements to adapt to specific environmental conditions and meet the functional requirements of specific tasks. Different modular robots can selectively be equipped with one or more of the above components according to preset task requirements. Through flexible component selection, the robot assembly can be reconfigured according to changing task requirements without modifying the hardware and software architecture, possessing strong generalization capabilities and breaking through the technical bottleneck of traditional robots with single functions.

[0013] The fault perception module includes a self-testing mechanism and a real-time monitoring mechanism, which is achieved by configuring specific types of fault detection sensors on the robot's task execution module: the motion execution component has a built-in vibration sensor to collect harmonic characteristics of mechanical transmission components. By analyzing abnormal changes in the vibration spectrum, potential mechanical hazards such as gear wear and bearing failure can be detected early; the vision component, motion component, and perception component have built-in current and voltage sensors to monitor power fluctuations and abnormal load conditions. By analyzing current and voltage waveform curves, electrical faults such as circuit short circuits and component overheating can be accurately identified; the types of sensors built into the dedicated components are customized according to the component's functional characteristics and application scenarios.

[0014] Before assembly begins, each modular robot uses the aforementioned fault detection sensors to execute an independent startup self-test program and sends a binary status code in the format of {hardware status code, sensor verification value} to the central monitoring and diagnostic subsystem. The startup self-test mechanism ensures that only fully functional modular robots are assembled into a robot assembly, thus blocking the possibility of faulty modules entering the system from the source.

[0015] After assembly, the fault detection sensors on the robot assembly monitor the working status of the corresponding components in real time and send a binary status code in the format of {hardware status code, sensor verification value} to the central monitoring and diagnostic subsystem at preset time intervals, establishing a component-level fault monitoring network. The preset time interval can be dynamically adjusted according to the criticality of the task, preferably within the range of 100-1000ms.

[0016] The present invention is further configured such that: the docking and communication module is implemented through a magnetic spring pin interface and a communication module supporting different communication protocols, preferably UWB, ZigBee, WiFi, and LoRa communication protocols;

[0017] The modular robots are physically connected via a magnetic spring pin interface, and data is transmitted between them after the connection is completed. After the assembly process is completed, the robot assembly and the central monitoring and diagnostic subsystem exchange data through a preset communication protocol. The communication protocol supports wired and / or wireless communication. The specific selection of the communication protocol and corresponding communication module depends on the power consumption requirements, bandwidth requirements and workspace size of the robot system.

[0018] The charging and power supply module is implemented through a lithium battery and a magnetic spring pin interface. Each modular robot has a built-in lithium battery for power supply and is charged through the magnetic spring pin interface.

[0019] The present invention is further configured such that: the multi-robotic arm collaborative subsystem is the core of the present invention, including at least two robotic arms that work collaboratively in a shared workspace to perform assembly or repair tasks;

[0020] For assembly tasks, the goal of each robotic arm is to work together to move different modular robots to designated locations within the workspace, ensuring precise docking of the magnetic spring pin interfaces of each module. For repair tasks, the goal of each robotic arm is to work together to remove the faulty modular robot and replace it with a functionally identical modular robot.

[0021] For dual-arm working scenarios, the assembly or repair process is divided into an environment modeling stage and a task execution stage: In the environment modeling stage, the monitoring module of the central monitoring and diagnostic subsystem models the modular robot and robotic arm, calculates the minimum distance vector between the two robotic arms in the shared workspace, discretizes the sampled robotic arm motion paths, and constructs a collision matrix; the high-dimensional configuration space is mapped to the coordinated sampling space and the coordinated time space respectively, thereby reducing the dimensionality of the high-dimensional configuration space.

[0022] During the task execution phase, an offline-online dual-layer planning architecture is adopted. Before the robotic arm moves, the monitoring module performs offline planning, using a trapezoidal velocity curve as the robotic arm motion control strategy to plan the time-optimal trajectory of each robotic arm. The parameter model of the trapezoidal velocity curve is as follows: Where ω mj Let t be the maximum velocity of the trapezoidal velocity curve of the j-th joint. f N represents the total motion time. jThe degrees of freedom of the robotic arm are defined as follows. Subsequently, the monitoring module performs collision detection on the generated trajectory and calculates the priority of the robotic arm's actions based on the collision detection results. The speed of the low-priority robotic arm's trajectory is adjusted, including linear deceleration and / or time delay, so that its trajectory avoids all potential collision areas. This effectively solves the path interference problem in the cooperative control of two robotic arms without increasing the planning complexity.

[0023] After the robotic arm starts moving, the monitoring module performs online planning, synchronizes the motion state parameters of each robotic arm, detects collision risks, adjusts the temporary target positions of low-priority robotic arms, and performs trajectory replanning through the local trajectory planner to quickly generate new conflict-free trajectories, ensuring the system operates safely in dynamic environments and improving the system's adaptability to environmental changes.

[0024] For multi-robotic arm work scenarios involving three or more arms, the monitoring module employs a pairwise coordination mechanism and a global coordination mechanism. The pairwise coordination mechanism will coordinate N... arm Each robotic arm is paired up to form a... Group of "dual robotic arms" N arm The number of robotic arms represents the number of tasks, transforming the "multi-robotic-arm task planning problem" into a "multi-pair task planning problem." Offline and online planning are performed in parallel to avoid local conflicts between pairs of robotic arms and generate priority suggestions. A global coordination mechanism performs global analysis and optimization of the priority relationships between each pair of robotic arms, employing a trajectory parameter decoupling algorithm to reduce computational complexity. Combined with the priority suggestions from the pair coordination mechanism, a global priority sequence is determined, avoiding conflicts between pairs of robotic arms at the system-wide level. The combination of a hierarchical processing mechanism and a parallel computing strategy enables the system to adapt to larger-scale multi-robotic-arm collaborative operations, efficiently achieving the assembly and repair of modular robots.

[0025] The present invention is further configured such that the central monitoring and diagnostic subsystem includes: a monitoring module for modeling calculation, visual monitoring, and manual intervention; and a diagnostic module for fault tracing, lifespan prediction, abnormal condition detection, and log recording.

[0026] The monitoring module models each link and joint of the robotic arm and the modular robot, and performs motion planning and calculates the robotic arm's trajectory based on the modeling results. It provides a real-time visualization of each robotic arm's operating status parameters, current task allocation, and overall task completion progress through an intuitive graphical user interface, enabling operators to clearly understand the system's operating status. It receives images from high-definition cameras deployed in the work environment, displaying the work site to the user in real time. Upon receiving a fault signal from the modular robot or robotic arm, or at any time required by the operator, the module allows for manual control of the robotic arm, as well as adding, modifying, pausing, and canceling assembly and repair tasks. This establishes an efficient human-machine collaboration mechanism, significantly improving the system's emergency response capabilities and processing efficiency under various complex working conditions.

[0027] The diagnostic module receives sensor data from each modular robot, uses pattern recognition algorithms to assess operational status and analyze fault sources, and identifies typical fault modes. It also receives data streams from cameras and robotic arm force sensors in the working environment, and analyzes the assembly or repair task execution process in real time, based on algorithm execution. When anomalies that the algorithm cannot automatically resolve are detected, including deadlock, robotic arm collision, task completion failure, infeasible task planning, or robotic arm task priority conflicts, a stop command is issued to the robotic arm, a fault status code is displayed, and manual intervention is requested. A lifespan prediction model is constructed through historical data analysis to predict the lifespan of different types of modular robots, and maintenance or replacement recommendations are generated periodically. Each assembly or repair operation is recorded, creating a traceable system log.

[0028] Through the collaborative work of the aforementioned monitoring and diagnostic modules, this invention achieves intelligent monitoring, accurate diagnosis, and proactive prevention throughout the entire system operation process, significantly improving the reliability and security of system operation.

[0029] The beneficial effects of this invention are as follows:

[0030] 1. This invention improves the fault detection capability of modular robot systems by setting up a self-checking mechanism. Before assembly, each modular robot performs a comprehensive self-check process, monitoring and reporting abnormal states in real time. This self-checking mechanism ensures that potentially faulty modular robots can be accurately identified and isolated before the assembly operation begins, fundamentally preventing faulty modular robots from being installed into the system, reducing the risk of task failure, and improving the reliability and stability of the system.

[0031] 2. The modular component architecture of this invention adopts a highly flexible and scalable design concept, supporting the selection of various action execution components, vision components, motion components, and perception components. It breaks through the limitations of traditional modular robot peripheral interfaces with limited and simple functions, greatly improving the robot's perception, motion, operation, interaction and other capabilities. This enables the system to quickly adapt to complex and ever-changing task environments through the free combination of components, achieving seamless switching of application scenarios.

[0032] 3. This invention achieves fully automated fault repair functionality for modular robots through multi-robotic arm collaborative technology. This not only reduces manpower consumption but also shortens system downtime and improves production efficiency. In high-risk, unattended, and remote operating environments, this autonomous repair capability is particularly important, laying a solid technical foundation for the widespread adoption of modular robot systems in broader and more extreme applications.

[0033] 4. To address the high computational complexity of traditional centralized control methods for multiple robotic arms, this invention proposes an offline-online dual-layer planning architecture, which effectively reduces the computational complexity of multi-arm collaborative methods. Furthermore, to address the limitations of traditional decoupled methods in adapting to collaborative scenarios involving three or more robotic arms and the potential for motion deadlock due to a lack of global coordination, this invention innovatively introduces pairwise coordination mechanisms and a global coordination mechanism to ensure efficient and orderly cooperation among three or more robotic arms in collaborative operations. Attached Figure Description

[0034] Figure 1 This is a schematic diagram of the system architecture of the present invention;

[0035] Figure 2 This is a detailed schematic diagram of the system architecture of the present invention;

[0036] Figure 3 This is a schematic diagram of the modular robot subsystem in this invention;

[0037] Figure 4 This is a detailed schematic diagram of the modular robot subsystem in this invention;

[0038] Figure 5 This is a schematic diagram of the multi-robotic arm collaborative subsystem in this invention;

[0039] Figure 6 This is a schematic diagram of the central monitoring and diagnostic subsystem in this invention;

[0040] Figure 7 This is a schematic diagram of the system architecture of a specific embodiment of the present invention;

[0041] Figure 8 This is a schematic diagram of the component library composition in a specific embodiment of the present invention; Detailed Implementation

[0042] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0043] Example 1

[0044] refer to Figure 1 and Figure 2 The modular robot assembly and repair system based on multi-manipulator collaboration provided by this invention includes a modular robot subsystem, a multi-manipulator collaboration subsystem, and a central monitoring and diagnostic subsystem. The architecture of the modular robot subsystem is shown below. Figure 3 and Figure 4 The architecture of the multi-robotic arm collaborative subsystem is shown in [link to architecture]. Figure 5 The architecture of the central monitoring and diagnostic subsystem is shown in [link to architecture]. Figure 6 To more clearly and completely describe the invention, this embodiment provides an implementation scheme applied to an unmanned research base in a desert. This scheme fully utilizes multi-robotic arm collaborative technology to achieve the assembly and repair functions of a modular robot.

[0045] refer to Figure 7 The system comprises: three high-precision six-DOF robotic arms 101, 102, and 103 located within the base, with a preferred positioning accuracy of ±0.05mm; a component library 104; a central monitoring and diagnostic subsystem 105; and two portable simplified robotic arms 106 and 107 and a simplified component library 108 deployed at a temporary maintenance point in the desert area. The central monitoring and diagnostic subsystem includes a monitoring module 105a and a diagnostic module 105b.

[0046] refer to Figure 8 Component library 104 and simplified component library 108 store various modular robots equipped with different components. Components include basic motion execution components 104a, vision components 104b, motion components 104c, and perception components 104d, as well as special components adapted to the special environment of the desert. Special components include: a sand-proof motion component 104e equipped with a high-power brushless motor, preferably with an IP68 protection rating, a fluororubber sealing ring material preferably used, and an operating temperature range preferably from -40℃ to +85℃; a special geological sampling motion execution component 104f, preferably with a sampling depth of up to 50cm and a sampling accuracy error ≤0.5mm; and a high-temperature resistant perception component 104g, preferably with an operating temperature range of -20℃ to +90℃ and a measurement accuracy preferably ≤±0.1℃.

[0047] To enable the modular robot to perform self-testing and real-time fault monitoring, the motion execution component incorporates a high-precision piezoelectric vibration sensor with an accuracy preferably above 0.01g, a sampling frequency preferably not less than 1kHz, and a power consumption fluctuation detection sensitivity preferably ≤5mW, thereby acquiring the harmonic characteristic spectrum of the mechanical transmission components in real time. The vision, motion, perception, and dedicated components incorporate current / voltage sensors with a measurement accuracy preferably above 0.1% and a data acquisition frequency preferably not less than 100Hz, thereby dynamically monitoring power fluctuations and abnormal loads.

[0048] The central monitoring and diagnostic subsystem adopts a high-performance computing platform, configured with a multi-core server-grade CPU, preferably with a main frequency of 3.0GHz or higher, and a multi-GPU acceleration array, preferably with a single card memory of 16GB or more, to meet the low latency and high throughput requirements of real-time working environment modeling and multi-robotic arm collaborative motion trajectory calculation.

[0049] The robot assembly's communication functionality is achieved through a communication module that supports both WiFi and LoRa protocols. In long-distance desert operations, considering the large workspace, the robot assembly communicates with the base using the LoRa protocol, which has a relatively long communication range.

[0050] The workflow of this embodiment is as follows:

[0051] First, the central monitoring and diagnostic subsystem 105, based on the pre-set desert scientific expedition requirements, calls upon a large language model to plan a modular robot assembly scheme and sends instructions to the three robotic arms 101, 102, and 103 within the base. The system then selects a modular robot with suitable task capabilities and performance parameters from the component library 104. The monitoring module 105a provides visual monitoring of the operating status of each robotic arm, the current task allocation, and the overall task completion progress. It receives images from high-definition cameras deployed in the working environment, providing users with a real-time, comprehensive view of the operational site. Upon receiving a fault code signal or a robotic arm abnormality report, or at any time required by the operator, the monitoring module allows the operator to intervene in the system's operation, enabling precise manual control of the robotic arms and the addition, modification, pause, and cancellation of tasks.

[0052] Three robotic arms begin assembling the robot assembly. Their objective is to collaboratively move different modular robots to designated locations within the workspace, ensuring precise docking of the magnetic spring-pin interfaces between modules. The process begins with environmental modeling. The monitoring module 105a of the central monitoring and diagnostic subsystem 105 models each link and joint of the robotic arms and the modular robots, calculating the minimum distance vector between the three robotic arms in the shared workspace in real time, discretizing the sampled robotic arm motion paths, and constructing a three-dimensional collision matrix. The monitoring module 105a maps the high-dimensional configuration space to a coordinated sampling space and a coordinated time space to reduce the dimensionality of the high-dimensional configuration space.

[0053] During the task execution phase, an offline-online dual-layer planning architecture is adopted. Before the robotic arms move, the monitoring module 105a first performs offline planning. The pair coordination mechanism transforms the "three-robotic arm task planning problem" into a "three-group dual-robotic arm parallel task planning problem," activating three pair coordination modules to handle potential collisions between the three groups of robotic arm pairs. Based on the collision matrix, the pair coordination mechanism performs collision detection on the generated trajectory and calculates the robotic arm action priority based on the collision detection results. Speed ​​adjustment is implemented on the motion trajectory of low-priority robotic arms, including linear deceleration and / or time delay, and priority suggestions are generated. After the three robotic arms begin collaborative operation, the monitoring module 105a enters the online planning phase. The robotic arms synchronize motion state parameters through real-time communication, with a communication cycle preferably below 10ms. If the monitoring module 105a detects a collision risk through the collision matrix, it adjusts the temporary target position of the low-priority robotic arm based on the collision matrix and the current motion state of the robotic arm, and performs trajectory replanning through a local trajectory planner.

[0054] During offline and online planning, the global coordination mechanism performs global analysis and optimization of the priority relationship of each group of robotic arms, adopts trajectory parameter decoupling algorithm to reduce computational complexity, and combines the priority suggestions given by the pair coordination mechanism to determine the global priority sequence. From the overall system level, conflicts between each group of dual robotic arms are avoided. Finally, the deadlock-free global priority sequence is determined as: robotic arm 102 > robotic arm 101 > robotic arm 103.

[0055] Three robotic arms successfully coordinated to assemble a detection robot assembly 201 equipped with a vision component 104b, a motion component 104c, a special geological sampling action execution component 104f, and a high-temperature resistant sensing component 104g.

[0056] After assembly, each modular robot constituting the exploration robot assembly 201 performs a self-test. The high-precision piezoelectric vibration sensor built into the special geological sampling action execution component 104f collects harmonic characteristics of the mechanical transmission components in real time; the current and voltage sensors built into the vision component 104b, motion component 104c, and high-temperature sensing component 104g dynamically monitor power fluctuations and abnormal loads. Each component sends a binary status code containing a hardware status code and sensor calibration values ​​to the central monitoring and diagnostic subsystem 105. The central monitoring and diagnostic subsystem 105 confirms that all component status codes are 0x01, indicating normal component status, and permits the exploration robot assembly 201 to leave the base and begin its desert exploration mission.

[0057] The robotic assembly 201 was performing tasks such as topographic mapping and soil sampling in a desert environment. During the operation, the robotic assembly suddenly and abnormally decelerated and was eventually forced to stop. The vibration sensor built into the motion component 104c detected abnormal harmonic spectrum during motor operation, indicating that sand and dust may have blocked the motor.

[0058] The detection robot assembly 201 sends a fault report to the base via the Lora protocol, which includes a detailed binary status code: {0x1E, 0.874}, where 0x1E is the abnormal hardware status code and 0.874 is the verification value of the vibration sensor, indicating that the vibration amplitude exceeds the preset threshold of 87.4%.

[0059] After receiving the fault report, the central monitoring and diagnostic subsystem 105 uses a pattern recognition algorithm to assess the operational status and analyze the fault source, basically determining that the fault is caused by dust clogging the motor and resulting in a failure of the motion components. Considering that the probe robot assembly 201 has lost its ability to move continuously, the system decides to assemble a traction robot assembly 202 to provide support.

[0060] The central monitoring and diagnostic subsystem 105 instructs three robotic arms 101, 102, and 103 to collaboratively assemble a traction robot assembly 202. This traction robot assembly has a dust-resistant motion component 104e equipped with a high-power brushless motor to enhance its adaptability to complex sandy terrain and its traction capacity, preferably with a maximum traction load of ≥30kg.

[0061] After the traction robot assembly 202 is assembled and completes its self-test, it proceeds to the location of the fault detection robot assembly 201. Upon arrival, the two robot assemblies dock using their respective pre-installed high-strength magnetic spring pin interfaces.

[0062] The traction robot assembly 202 pulls the fault detection robot assembly 201 back to the temporary repair point in the desert area. The two simplified robotic arm systems 106 and 107 at the temporary repair point receive repair instructions from the central monitoring and diagnostic subsystem 105. Their task is to work together to remove the faulty motion component 104c and replace it with a sand-proof motion component 104e equipped with a high-power motor.

[0063] Two simplified robotic arms, 106 and 107, began the repair operation, successively completing environmental modeling and task execution. Because there are only two robotic arms, the system employs a dual-arm collaborative algorithm, eliminating the need for pairwise or global coordination mechanisms. Robotic arms 106 and 107 successfully collaborated to disassemble motion component 104c and replaced it with a spare, dust-proof motion component 104e equipped with a high-power motor from the simplified component library 108. This upgraded component features increased motor power, optimized motor heat dissipation design, and an added dust-proof sealing structure.

[0064] After installation, the exploration robot assembly 201 performs a self-test, with each component sending a status code to the central monitoring and diagnostic subsystem 105. The newly installed reinforced sand and dust resistant motion component 104e has a status code of 0x01, indicating normal operation and successful repair. The exploration robot assembly 201 resumes normal function and continues its desert exploration mission. The entire process is automated and requires no human intervention.

[0065] Meanwhile, the diagnostic module 105b of the central monitoring and diagnostic subsystem 105 performs multi-dimensional service life analysis on the faulty moving components and updates the service life prediction model for these components in desert environments using a time series prediction model. The preferred time series prediction model is an integrated model combining gradient boosting decision trees and long short-term memory networks. The diagnostic module 105b records the assembly and repair operations, constructing a traceable system log.

[0066] When using this invention, multiple technical implementation paths can be selected according to specific scenario requirements. Specific alternative solutions are explained below: First, regarding the trapezoidal velocity curve control strategy for the robotic arm adopted in this invention, the following alternative solutions can be flexibly selected according to application requirements: The S-shaped velocity curve planning strategy, by introducing acceleration smoothing constraints, achieves continuous controllability of the acceleration change rate, effectively reducing structural vibration and impact load on the mechanical system, extending equipment lifespan, and is particularly suitable for high-precision operation scenarios; trajectory planning based on spline interpolation can generate smooth trajectories with continuous high-order derivative characteristics, allowing the robotic arm to achieve more natural motion transitions between key points while maintaining global continuity of velocity and acceleration, suitable for assembly tasks with high motion precision requirements; path planning based on the artificial potential field method treats obstacles as repulsive force fields and target points as attractive force fields, automatically constructing obstacle avoidance trajectories in complex environments, reducing the risk of robotic arm collisions. Each of the above alternative solutions has its own technical advantages and can be flexibly selected and implemented according to specific application scenario requirements and system configuration.

[0067] Secondly, the core design concept of this invention lies in the highly modular design of its components. Whether it's motion execution components, vision components, motion components, perception components, or specialized components, all can be flexibly replaced or expanded according to specific task requirements after standardized hardware interface adaptation is completed. For example, motion execution components may include mechanical grippers, electromagnetic chucks, pneumatic push rods, and soft grippers; vision components may include lidar, infrared cameras, and binocular depth cameras; motion components may include tracked chassis and four-wheel omnidirectional wheel chassis; perception components may include gas sensors, metal detectors, and multispectral ranging sensors; and specialized components also follow a modular design logic, and may include deep-sea pressure-resistant motion components and polar cryogenic ice drilling components.

[0068] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A modular robot assembly and repair system based on multi-robotic arm collaboration, characterized in that, include: Modular robot subsystems, acting as objects manipulated by robotic arms, dock to form robot assemblies; Multi-arm collaborative subsystem for robot assembly and repair operations; The central monitoring and diagnostic subsystem is used for system status monitoring and diagnostics.

2. The modular robot assembly and repair system based on multi-robotic arm collaboration according to claim 1, characterized in that, The modular robot subsystem includes: a task execution module for the robot to complete its assigned tasks; a fault perception module for detecting whether the robot has malfunctioned; a docking and communication module for communication and docking between the robot and the modular robot; and a charging and power supply module for charging and powering the robot.

3. The modular robot assembly and repair system based on multi-robotic arm collaboration according to claim 2, characterized in that, The task execution module includes optional motion execution components, vision components, motion components, perception components, and special-purpose components: the motion execution components include end effectors for performing various operational tasks; the vision components include vision acquisition and processing instruments for performing visual tasks; the motion components include a chassis system for performing various motion tasks; the perception components include sensors for assisting the robot in perceiving its own state and the surrounding environment; and the special-purpose components include components designed for specific working environments or task requirements to adapt to specific environmental conditions and meet the functional requirements of specific tasks. Different modular robots can selectively be equipped with one or more of the above components according to preset task requirements. The fault perception module includes a self-testing mechanism and a real-time monitoring mechanism, which is achieved by configuring specific types of fault detection sensors on the robot's task execution module: the motion execution component has a built-in vibration sensor to collect harmonic characteristics of mechanical transmission components; the vision component, motion component, and perception component have built-in current and voltage sensors to monitor power fluctuations and abnormal load conditions; the types of sensors built into the dedicated components are customized according to the component's functional characteristics and application scenarios. The docking and communication module is implemented through a magnetic spring pin interface and a communication module that supports different communication protocols: each modular robot is physically docked with the other using a magnetic spring pin interface, and data is transmitted between them through this interface after docking is completed; after the assembly process is completed, data is exchanged between the robot assemblies and between the robot assemblies and the central monitoring and diagnostic subsystem through a preset communication protocol, which supports wired and / or wireless communication. The charging and power supply module is implemented through a lithium battery and a magnetic spring pin interface. Each modular robot has a built-in lithium battery for power supply and is charged through the magnetic spring pin interface.

4. The modular robot assembly and repair system based on multi-robotic arm collaboration according to claim 1, characterized in that, The multi-robotic arm collaborative subsystem includes at least two robotic arms that work collaboratively in a shared workspace to perform assembly or repair tasks. For assembly tasks, the goal of each robotic arm is to work together to move different modular robots to designated locations within the workspace, ensuring precise docking of the magnetic spring pin interfaces of each module. For repair tasks, the goal of each robotic arm is to work together to remove the faulty modular robot and replace it with a functionally identical modular robot. For dual-arm working scenarios, the assembly or repair process is divided into an environment modeling stage and a task execution stage: In the environment modeling stage, the monitoring module of the central monitoring and diagnostic subsystem models the modular robot and robotic arm, calculates the minimum distance vector between the two robotic arms in the shared workspace, discretizes the sampled robotic arm motion paths, and constructs a collision matrix; the high-dimensional configuration space is mapped to the coordinated sampling space and the coordinated time space respectively, thereby reducing the dimensionality of the high-dimensional configuration space. During the task execution phase, an offline-online dual-layer planning architecture is adopted. Before the robotic arm moves, the monitoring module performs offline planning, plans the optimal time trajectory for each robotic arm, performs collision detection on the generated trajectory, and calculates the priority of robotic arm actions based on the collision detection results. The speed adjustment is implemented on the motion trajectory of low-priority robotic arms, including linear deceleration and / or time delay, so that their motion trajectory avoids all potential collision areas. After the robotic arm starts moving, the monitoring module performs online planning, synchronizes the motion state parameters of each robotic arm, detects collision risks, adjusts the temporary target positions of low-priority robotic arms, and performs trajectory replanning through a local trajectory planner to quickly generate a new, conflict-free trajectory.

5. A modular robot assembly and repair system based on multi-robotic arm collaboration according to claim 4, characterized in that, For multi-robotic arm work scenarios involving three or more arms, the monitoring module employs a pairwise coordination mechanism and a global coordination mechanism. The pairwise coordination mechanism will coordinate N... arm Each robotic arm is paired up to form a... The "pair of two robotic arms" transforms the "multi-robotic arm task planning problem" into a "multiple pairs of two robotic arm task planning problem," performing offline and online planning in parallel to avoid local conflicts between each pair of robotic arms and generate priority suggestions. A global coordination mechanism performs global analysis and optimization of the priority relationships between each pair of robotic arms, uses a trajectory parameter decoupling algorithm to reduce computational complexity, and combines the priority suggestions given by the pair coordination mechanism to determine the global priority sequence, thus avoiding conflicts between pairs of two robotic arms at the overall system level.

6. The modular robot assembly and repair system based on multi-robotic arm collaboration according to claim 1, characterized in that, The central monitoring and diagnostic subsystem includes: a monitoring module for modeling calculations, visual monitoring, and manual intervention; and a diagnostic module for fault tracing, lifespan prediction, abnormal condition detection, and log recording. The monitoring module models each link and joint of the robotic arm and the modular robot, performs robotic arm motion planning, and calculates the robotic arm's motion trajectory; it visualizes the operating status parameters, current task allocation, and total task completion progress of each robotic arm; it receives images from high-definition cameras deployed in the working environment and displays the work site; when a fault signal is received from the modular robot or robotic arm, or at any time required by the operator, the operator is allowed to intervene in the system operation to achieve manual control of the robotic arm, as well as add, modify, pause, and cancel assembly and repair tasks; The diagnostic module receives sensor data from each modular robot to assess its operational status and analyze fault tracing; it also receives data from cameras and robotic arm force sensors in the working environment to analyze the assembly or repair process in real time. When an abnormal situation that the algorithm cannot automatically resolve is detected, a stop command is issued to the robotic arm, a fault status code is displayed, and manual intervention is requested. A service life prediction model is constructed through historical data analysis to predict the service life of different types of modular robots and to generate maintenance or replacement recommendations periodically. Each assembly or repair operation is recorded to build a traceable system log.

Images