Multifunctional mechanical arm system for distribution network unmanned aerial vehicle inspection

Abstract

The invention relates to the field of unmanned aerial vehicle inspection, in particular to a multifunctional mechanical arm system for distribution network unmanned aerial vehicle inspection. The system comprises an unmanned aerial vehicle platform, a mechanical arm module, a bionic control system and a replaceable operation assembly. The unmanned aerial vehicle platform provides flight control and load bearing functions; the mechanical arm module executes grabbing, spraying and laser operation through a multi-degree-of-freedom joint structure and a driving unit; the bionic control system collects arm motion trail data of an operator in real time and converts the arm motion trail data into a mechanical arm motion instruction, and synchronous motion control is achieved. The replaceable operation assembly is connected through a standardized interface, and different operation requirements are met. Through the bionic control system, the problems of complex operation and insufficient precision in the prior art are solved, and the precision and response speed of the unmanned aerial vehicle in tasks such as laser foreign matter removal, insulating paint spraying and foreign matter grabbing are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the field of UAV inspection, and particularly to a multi-functional robotic arm system for distribution network UAV inspection. Background Art

[0002] With the rapid development of UAV technology, its application in the field of distribution network inspection has gradually attracted attention. Traditional UAV inspection systems mainly rely on simple cameras or sensors for remote monitoring, but such systems have many limitations when performing complex tasks. For example, existing UAV inspection systems often face problems such as insufficient accuracy and slow response speed when performing operations such as laser foreign object removal, insulating paint spraying, and foreign object grasping. In addition, the control methods of traditional UAVs are mostly joystick operations, which are complex to operate and require high skills from operators, making it difficult to achieve high-precision operations.

[0003] To overcome these deficiencies, some studies have proposed improvement schemes. For example, Patent CN116079691A discloses a UAV-mounted robotic arm and its intelligent control system, which realizes precise clamping of target items through the opening and closing and rotation functions of the robotic arm. However, this scheme still fails to solve the problems of operation complexity and insufficient operation accuracy. In addition, although there are schemes in the prior art that use myoelectric signals to control robotic arms, they are mainly applied to the field of rehabilitation robots and are not combined with UAV platforms for application.

[0004] The present invention addresses the problems existing in the prior art and proposes a multi-functional robotic arm system for distribution network UAV inspection. Summary of the Invention

[0005] The purpose of the present invention is to provide a multi-functional robotic arm system for distribution network UAV inspection to solve the problems of operation complexity and insufficient operation accuracy in the prior art.

[0006] To achieve the above purpose, the following technical solutions are adopted.

[0007] A multi-functional robotic arm system for distribution network UAV inspection includes:

[0008] A UAV platform for providing flight control and load-bearing functions;

[0009] A robotic arm module detachably connected to the lower part of the UAV platform, including a multi-degree-of-freedom joint structure and a driving unit, and the robotic arm module is used to perform grasping, spraying, and laser operations;

[0010] The bionic control system includes a motion capture unit and an instruction conversion unit. The motion capture unit collects the motion trajectory data of the operator's arm in real time, and converts the motion trajectory data into the action instructions of the robotic arm module through the instruction conversion unit, realizing the synchronous motion control of the robotic arm module and the operator's arm;

[0011] The replaceable operation component includes a variety of operation components, which are connected to the end of the robotic arm module through a standardized interface.

[0012] Optionally, the bionic control system further includes:

[0013] An electromyogram signal analysis unit for collecting the electromyogram signals of the operator's arm through the bioelectric sensors of the wearable device, and performing time-domain noise suppression processing on the electromyogram signals to generate a denoised electromyogram waveform;

[0014] A motion trajectory reconstruction unit for inputting the denoised electromyogram waveform into a pre-trained neural network model and outputting the three-dimensional motion trajectory vector of the operator's arm. The neural network model is trained based on historical operation data and is used to eliminate signal distortion caused by environmental electromagnetic interference;

[0015] A joint dynamic mapping unit for generating the rotation angle, angular velocity and torque instructions of each joint according to the inverse solution mapping relationship between the three-dimensional motion trajectory vector and the robotic arm joint kinematic model, and transmitting them to the main control unit of the UAV platform through a wireless communication module;

[0016] A delay compensation module for predicting the instruction transmission delay based on the wireless communication link state, and preloading the action instructions in the robotic arm joint controller through a feedforward control algorithm to eliminate operation lag.

[0017] Optionally, the joint structure of the robotic arm module includes:

[0018] A harmonic reduction rotating joint for providing continuous rotational motion in the horizontal plane through a strain wave gear. A multi-axis torque sensor is integrated at the end of the harmonic reducer to monitor the rotational load torque in real time and feedback it to the dynamic drive unit;

[0019] A planetary roller pitching joint for realizing 0°-90° pitching swing through a roller screw pair. A vibration damping layer is provided at the end of the screw to suppress the positioning error caused by high-frequency vibration;

[0020] A magnetic levitation telescopic joint for driving the telescopic arm to move along the axis through a linear motor. A non-contact Hall displacement sensor is integrated at the end of the telescopic arm to detect the displacement in real time and feedback it to the dynamic drive unit;

[0021] A temperature-torque coupling monitoring unit is used to collect the correlation data of the motor temperature and the output torque of each joint in real time. When abnormal temperature or torque mutation is detected, it triggers a frequency reduction protection mechanism and generates an alarm signal.

[0022] Optionally, the modular interface of the replaceable operation component includes:

[0023] A multi-modal positioning mechanism is used to work in cooperation with a lidar and a near-field communication chip to obtain the three-dimensional attitude data of the operation component during component installation;

[0024] An active compensation unit is used to drive a micro servo motor to adjust the position of the interface pin based on the three-dimensional attitude data, so that the installation angle error of the component is less than ±0.5°, and triggers an electrical connection self-check program to verify the contact reliability;

[0025] A redundant locking unit is used to trigger a secondary locking action through a shape memory alloy after magnetic attraction positioning and mechanical buckling are completed, so as to ensure the anti-vibration stability during high-altitude operation.

[0026] Optionally, the drive unit further includes:

[0027] A load prediction module is used to analyze historical load data through a long short-term memory network to predict the torque requirements of each joint within the next 5 seconds;

[0028] A dynamic allocation module is used to allocate motor power in advance based on the prediction results and optimize the start-stop timing of the energy recovery circuit;

[0029] An overload protection circuit is used to cut off the power supply of the motor and activate the emergency braking device of the robotic arm when it detects that the instantaneous current exceeds the safety threshold;

[0030] A bidirectional energy management unit is used to preferentially use battery power during the acceleration stage of the robotic arm and store the kinetic energy recovered into electrical energy in the drone battery during the deceleration stage to achieve dynamic energy balance.

[0031] Optionally, the obstacle avoidance module includes:

[0032] A multi-source perception fusion unit is used to construct a dynamic environment model within the range of 0.5m - 3m around the robotic arm through a millimeter-wave radar, an infrared thermal imager, and a depth camera;

[0033] A hierarchical decision-making unit is used to select one of the following operations according to the movement speed and intrusion direction of the obstacle:

[0034] When the intrusion speed is lower than the threshold, trigger a path replanning algorithm to generate an obstacle avoidance path;

[0035] When the intrusion speed is higher than the threshold, activate the electromagnetic brake to lock each joint and send a graded warning signal to the operator through the tactile feedback unit;

[0036] A self-learning optimization unit for training a reinforcement learning model based on historical obstacle avoidance data and dynamically adjusting the path planning strategy to improve the obstacle avoidance efficiency.

[0037] Optionally, the task planning module further includes:

[0038] A power grid topology analysis unit for importing 3D point cloud data of the distribution network line, identifying the spatial coordinates and defect levels of insulators, wire joints, and lightning arresters;

[0039] A multi-objective optimization unit for generating an inspection sequence that takes into account both safety and efficiency by combining the kinematic constraints of the robotic arm, the endurance time of the UAV, and the priority of equipment defects;

[0040] A cooperative control unit for decomposing the inspection sequence into the flight trajectory of the UAV and the action instructions of the robotic arm, and real-time correcting the vibration amplitude at the end of the robotic arm through a Kalman filter;

[0041] A dynamic priority adjustment unit for reallocating the operation sequence and updating the flight path when a sudden equipment failure is detected.

[0042] Optionally, the grasping fixture includes:

[0043] A multi-modal tactile perception unit for detecting the shape, hardness, and surface material distribution of the target object through a piezoresistive sensor array and a capacitive sensor;

[0044] A fuzzy adaptive clamping unit for dynamically adjusting the opening and closing angle of the jaws and the clamping force according to the tactile data, so that the contact pressure distribution is uniform and the peak pressure is less than the preset threshold;

[0045] A dual-mode insulation monitoring unit for working in parallel through an electric field sensor and a leakage current detection circuit, and triggering safety locking when the detected potential difference exceeds the voltage threshold or the leakage current exceeds the current threshold;

[0046] A self-cleaning mechanism for removing foreign objects on the contact surface through high-pressure air flow when the jaws are closed, ensuring the effective isolation of the insulation protection layer.

[0047] Optionally, the spraying device includes:

[0048] An ultrasonic viscosity detection unit for real-time monitoring of the rheological properties of the paint through a sound velocity sensor and generating a viscosity-temperature correlation curve;

[0049] A dynamic atomizing nozzle is used to adjust the nozzle aperture and atomizing air pressure according to the flight speed of the unmanned aerial vehicle, the curvature of the target surface, and the viscosity-temperature curve through a PID algorithm, so as to make the coating deposition uniformity reach more than 90%;

[0050] A coating thickness feedback unit is used to detect the spraying thickness in real time through a laser interferometer and trigger a supplementary spraying program when the thickness deviation exceeds the thickness threshold;

[0051] A self-repairing nozzle head is used to remove deposits and restore atomizing performance through reverse pulse air flow when a nozzle blockage is detected.

[0052] Optionally, the laser processing unit includes:

[0053] A spectral material identification module is used to analyze the composition of the target foreign object through reflection spectroscopy and match the laser wavelength-power combination table in a preset database;

[0054] A dynamic focusing module is used to adjust the beam divergence angle in real time according to the distance of the foreign object, so as to control the energy density error within a preset range;

[0055] A safety interlock unit is used to reduce the laser power to a safe level when a person is detected to enter the operation radius of 2 m, and generate an electronic fence warning signal to be linked and locked with a physical shielding cover;

[0056] An energy distribution optimization module is used to optimize the laser scanning path based on the foreign object material and thickness distribution through a genetic algorithm to improve the cleaning efficiency.

[0057] Compared with the prior art, the present invention has the following beneficial effects:

[0058] The present invention proposes a multi-functional robotic arm system for power distribution unmanned aerial vehicle inspection, which realizes the synchronous motion control of the robotic arm and the operator's arm through a bionic control system. Effectively solves the problems of insufficient operation complexity and operation accuracy in the prior art, and significantly improves the accuracy and response speed of the unmanned aerial vehicle when performing tasks such as laser foreign object removal, insulating coating spraying, and foreign object grasping. In addition, the lightweight design and low-power consumption system of the system further improve the mobility and operation stability of the unmanned aerial vehicle.

[0059] Through the electromyogram signal analysis unit and the motion trajectory reconstruction unit, the accuracy and stability of the bionic control system are further optimized, and the signal distortion caused by environmental electromagnetic interference can be effectively eliminated; the mechanical arm joint structure design improves the motion accuracy and load capacity of the mechanical arm; the modular interface of the replaceable operation component enhances the versatility and reliability of the system; the optimization of the drive unit further reduces the system power consumption and improves the energy utilization efficiency; the obstacle avoidance module design enhances the safety and adaptability of the system; the task planning module optimizes the inspection path and operation efficiency; the design of the grasping fixture, spraying device and laser processing unit provides efficient solutions for different operation tasks respectively. Brief Description of the Drawings

[0060] Figure 1 It is a schematic diagram of the modules of an embodiment of the multi-functional robotic arm system for distribution network UAV inspection according to the present invention. Detailed Embodiment

[0061] The present invention will be described in detail below with reference to the drawings and in conjunction with the embodiments. It should be noted that, without conflict, the embodiments in the present application and the features in the embodiments may be combined with each other.

[0062] The following detailed descriptions are all exemplary descriptions, aiming to provide further detailed descriptions of the present invention. Unless otherwise specified, all technical terms adopted by the present invention have the same meaning as commonly understood by those of ordinary skill in the art to which this application belongs. The terms used in the present invention are only for the purpose of describing specific embodiments, and are not intended to limit the exemplary embodiments according to the present invention.

[0063] As Figure 1 shown, the multi-functional robotic arm system for distribution network UAV inspection of the present invention is a highly integrated innovative system, and its design fully considers the flight performance of the UAV platform, the versatility of the robotic arm module, the high-precision operation of the bionic control system, and the flexibility of the replaceable operation components.

[0064] The UAV platform is the bearing foundation of the entire system, and its main function is to provide stable flight control and load-bearing capacity. The UAV platform usually adopts a multi-rotor structure, which has excellent flight performance of vertical takeoff and landing and hovering in the air, and does not require an airport, and can make use of the roofs or water surfaces of urban buildings for takeoff and landing. The energy utilization rate of the multi-rotor UAV is more than 30% higher than that of traditional helicopters, that is, the dynamic efficiency is more than 30% higher. In addition, the flight control and management subsystem of the UAV platform is the core system for completing the entire flight process such as takeoff, in-air flight, task execution, and return and landing. By controlling various sensors, navigation devices, actuators, etc. of the UAV, it ensures that the UAV can fly according to the preset route and complete various tasks.

[0065] The design of the drone platform needs to consider multiple aspects. First of all, its flight control system needs to have high-precision positioning and navigation capabilities. This can be achieved by integrating the Global Positioning System (GPS), Inertial Measurement Unit (IMU), and vision sensors. GPS provides accurate geographical location information, IMU is used to measure the acceleration and angular velocity of the drone, and vision sensors can be used for auxiliary navigation and obstacle avoidance. The data from these sensors are fused and processed by the flight controller to achieve precise flight control.

[0066] The energy system of the drone platform is also a crucial part. To meet the long-term inspection tasks, the drone needs to be equipped with a high-performance battery system. The selection of the battery needs to consider factors such as energy density, discharge rate, and cycle life. For example, lithium-ion batteries are widely used in drones due to their high energy density and long cycle life. In addition, the drone can also be equipped with solar panels to achieve longer endurance.

[0067] The structural design of the drone platform needs to consider the requirements of lightweight and high strength. This can be achieved by adopting advanced materials and manufacturing technologies. For example, carbon fiber composite materials are widely used in the structural components of drones due to their high strength and lightweight characteristics. In addition, the frame design of the drone also needs to consider aerodynamic performance to reduce flight resistance and improve energy utilization efficiency.

[0068] The drone platform also needs to have good communication capabilities. This includes communication with the ground control station and communication with other drones. The communication system needs to have high bandwidth, low latency, and anti-interference capabilities. For example, wireless communication technologies such as Wi-Fi, 4G / 5G, or dedicated frequency band communication can be used to ensure stable data transmission.

[0069] The robotic arm module is connected to the lower part of the drone platform in a detachable manner, which facilitates quick replacement and maintenance. The robotic arm module adopts a multi-degree-of-freedom joint structure, and each joint is equipped with a high-precision drive unit, which can achieve complex motion control. The joint design of the robotic arm can refer to existing collaborative robotic arm technologies. For example, an integrated joint design is adopted, which integrates components such as frameless torque motors, harmonic reducers, and encoders. The frameless torque motor retains the part of the traditional motor used to generate torque and speed, while removing the shaft, bearings, housing, or end cap, and is a frameless permanent magnet motor measured by output torque. In addition, various operation components such as gripper fixtures, spraying devices, and laser processing units can be connected to the end of the robotic arm module, and quick replacement can be achieved through standardized interfaces to meet different operation requirements.

[0070] The design of the robotic arm module needs to consider multiple aspects. First of all, its multi-degree-of-freedom joint structure needs to have high-precision motion control capabilities. This can be achieved by using high-precision encoders and drivers. The encoder is used to measure the angular position of the joint, while the driver is used to control the movement of the joint. In addition, the joint design of the robotic arm also needs to consider torque output and speed control. For example, a harmonic reducer can be used to increase torque output while maintaining a small volume and weight.

[0071] The end effector of the robotic arm module needs to be versatile. This can be achieved by designing replaceable working components. For example, a gripper can be used to grasp objects, a spraying device can be used for spraying operations, and a laser processing unit can be used for laser cutting or welding. These working components are connected to the robotic arm module through standardized interfaces, facilitating quick replacement and maintenance.

[0072] The structural design of the robotic arm module needs to consider the requirements of lightweight and high strength. This can be achieved by using advanced materials and manufacturing technologies. For example, aluminum alloys and carbon fiber composite materials are widely used in the structural components of the robotic arm due to their high strength and lightweight characteristics. In addition, the design of the robotic arm also needs to consider its motion range and load capacity. For example, the arm span of the robotic arm can be designed to be 2 meters to meet different working requirements, and its load capacity can be designed to not exceed 15 kg to ensure the mobility and operational stability of the drone.

[0073] The robotic arm module also needs to have good control capabilities. This can be achieved by using advanced control algorithms. For example, model-based control algorithms such as PID control or adaptive control can be used to achieve precise motion control. In addition, the robotic arm module can also be equipped with sensors such as force sensors and displacement sensors to achieve force feedback and displacement feedback control.

[0074] The bionic control system is one of the core innovations of the present invention, and its design inspiration comes from the motion control mechanism of organisms. The system includes a motion capture unit and an instruction conversion unit, which can collect the motion trajectory data of the operator's arm in real time and convert this data into action instructions for the robotic arm module. This control method significantly improves the operation accuracy and response speed, enabling the robotic arm to achieve synchronous motion control with the operator's arm. Further, the bionic control system also includes an electromyogram signal analysis unit, a motion trajectory reconstruction unit, a joint dynamic mapping unit, and a delay compensation module. The electromyogram signal analysis unit collects the electromyogram signals of the operator's arm through the bioelectric sensors of the wearable device and performs time-domain noise suppression processing. The motion trajectory reconstruction unit uses a pre-trained neural network model to convert the denoised electromyogram waveform into a three-dimensional motion trajectory vector of the operator's arm. The joint dynamic mapping unit generates the rotation angle, angular velocity, and torque instructions of each joint according to the inverse mapping relationship between the motion trajectory vector and the robotic arm joint kinematic model. The delay compensation module eliminates the operation lag by predicting the wireless communication link state and preloading the action instructions.

[0075] The design of the bionic control system needs to consider multiple aspects. First, its motion capture unit needs to have high-precision motion trajectory acquisition capabilities. This can be achieved by adopting advanced sensor technologies. For example, an optical motion capture system can be used. By installing marker points on the operator's arm, multiple cameras are used to capture the positions and motion trajectories of the marker points. In addition, the motion capture unit can also adopt inertial sensors, such as accelerometers and gyroscopes, to measure the acceleration and angular velocity of the operator's arm.

[0076] The instruction conversion unit needs to have efficient data processing capabilities. This can be achieved by adopting high-performance processors and algorithms. For example, an algorithm based on neural networks can be used to convert the collected motion trajectory data into action instructions for the robotic arm. The neural network model can be trained with a large amount of training data to improve its accuracy and robustness.

[0077] The electromyogram signal analysis unit needs to have high-precision electromyogram signal acquisition and processing capabilities. This can be achieved by adopting bioelectric sensors and signal processing algorithms. For example, the bioelectric sensors can collect the electromyogram signals of the operator's arm, and the signal processing algorithms can perform noise suppression and feature extraction on the electromyogram signals. Noise suppression can be achieved through time-domain filtering or frequency-domain filtering, and feature extraction can be achieved by extracting the time-domain features or frequency-domain features of the electromyogram signals.

[0078] The delay compensation module needs to have efficient operation delay compensation capabilities. This can be achieved by adopting advanced control algorithms. For example, a feedforward control algorithm can be used to preload action instructions by predicting the state of the wireless communication link to eliminate operation lags. In addition, the delay compensation module can also adopt a feedback control algorithm to dynamically adjust the loading time of action instructions by monitoring the operation delay in real time.

[0079] The design of the replaceable operation component takes into account modularity and versatility. The modular interface includes a multi-modal positioning mechanism, an active compensation unit, and a redundant locking unit. The multi-modal positioning mechanism works in cooperation with a lidar and a near-field communication chip to obtain the three-dimensional attitude data of the operation component during component installation. The active compensation unit drives a micro servo motor to adjust the position of the interface pins based on the three-dimensional attitude data, ensuring that the installation angle error of the component is less than ±0.5°, and triggering an electrical connection self-check program to verify contact reliability. After magnetic attraction positioning and mechanical buckling are completed, the redundant locking unit triggers a secondary locking action through a shape memory alloy to ensure anti-vibration stability during high-altitude operations.

[0080] The design of the replaceable operation component needs to consider multiple aspects. First of all, its modular interface needs to have high-precision positioning and connection capabilities. This can be achieved by adopting advanced sensor and driver technologies. For example, the multi-modal positioning mechanism can use a lidar and a near-field communication chip to obtain the three-dimensional attitude data of the operation component through cooperation. The lidar can measure the position and distance of the operation component, while the near-field communication chip can be used for data transmission and communication.

[0081] The active compensation unit needs to have efficient angle adjustment capabilities. This can be achieved by adopting a micro servo motor and a control algorithm. For example, the micro servo motor can drive the adjustment of the interface pin position according to the three-dimensional attitude data, ensuring that the installation angle error of the component is less than ±0.5°. In addition, the active compensation unit can also be equipped with an electrical connection self-check program to verify contact reliability.

[0082] The redundant locking unit needs to have highly reliable locking capabilities. This can be achieved by adopting shape memory alloy and mechanical buckling technologies. For example, the shape memory alloy can trigger a secondary locking action after magnetic attraction positioning is completed to ensure anti-vibration stability during high-altitude operations. In addition, the redundant locking unit can also be equipped with a safety detection mechanism to monitor the locking state in real time.

[0083] The design of the replaceable operation component also needs to consider its versatility and compatibility. This can be achieved by adopting a standardized interface and modular design. For example, the operation component can be connected to the robotic arm module through a standardized interface, facilitating quick replacement and maintenance. In addition, the design of the operation component also needs to consider its functional diversity and adaptability to meet different operation requirements.

[0084] The design of the drive unit takes into account efficient energy management and system stability. The load prediction module analyzes historical load data through a long short-term memory network to predict the torque requirements of each joint within the next 5 seconds. The dynamic allocation module allocates motor power in advance based on the prediction results and optimizes the start-stop timing of the energy recovery circuit. When the overload protection circuit detects that the instantaneous current exceeds the safety threshold, it cuts off the power supply to the motor and activates the emergency braking device of the robotic arm. The bidirectional energy management unit preferentially uses battery power during the acceleration phase of the robotic arm and recovers kinetic energy into electrical energy and stores it in the drone battery during the deceleration phase to achieve dynamic energy balance.

[0085] The design of the drive unit needs to consider multiple aspects. First of all, its load prediction module needs to have efficient data analysis capabilities. This can be achieved by adopting a long short-term memory network (LSTM). LSTM is a special type of recurrent neural network that can process and predict time series data. By analyzing historical load data, LSTM can predict the torque requirements of each joint within the next 5 seconds, thus achieving efficient load prediction.

[0086] The dynamic allocation module needs to have efficient energy allocation capabilities. This can be achieved by adopting advanced control algorithms and energy recovery circuits. For example, the dynamic allocation module can allocate motor power in advance according to the prediction results and optimize the start-stop timing of the energy recovery circuit. The energy recovery circuit can recover kinetic energy into electrical energy during the deceleration phase of the robotic arm and store it in the drone battery, thus achieving dynamic energy balance.

[0087] The overload protection circuit needs to have efficient safety protection capabilities. This can be achieved by adopting advanced circuit designs and control algorithms. For example, when the overload protection circuit detects that the instantaneous current exceeds the safety threshold, it cuts off the power supply to the motor and activates the emergency braking device of the robotic arm. In addition, the overload protection circuit can be equipped with a real-time monitoring mechanism to ensure the safe operation of the system.

[0088] The bidirectional energy management unit needs to have efficient energy management capabilities. This can be achieved by adopting advanced control algorithms and battery management systems. For example, the bidirectional energy management unit can preferentially use battery power during the acceleration phase of the robotic arm, recover kinetic energy into electrical energy during the deceleration phase, and store it in the drone battery. In addition, the bidirectional energy management unit can be equipped with an intelligent charging mechanism to optimize the service life of the battery.

[0089] The obstacle avoidance module achieves efficient obstacle avoidance through a multi-source perception fusion unit, a hierarchical decision-making unit, and a self-learning optimization unit. The multi-source perception fusion unit uses a millimeter-wave radar, an infrared thermal imager, and a depth camera to construct a dynamic environment model within a range of 0.5m - 3m around the robotic arm. The hierarchical decision-making unit selects an obstacle avoidance strategy based on the movement speed and intrusion direction of the obstacle. The self-learning optimization unit trains a reinforcement learning model based on historical obstacle avoidance data and dynamically adjusts the path planning strategy to improve the obstacle avoidance efficiency.

[0090] The design of the obstacle avoidance module needs to consider multiple aspects. First of all, its multi-source perception fusion unit needs to have efficient data acquisition and processing capabilities. This can be achieved by adopting multiple sensor technologies. For example, a millimeter-wave radar can be used to measure the distance and speed of obstacles, an infrared thermal imager can be used to detect the temperature distribution of obstacles, and a depth camera can be used to obtain the three-dimensional shape and position of obstacles. Through the fusion and processing of the data from these sensors, a dynamic environment model around the robotic arm can be constructed.

[0091] The hierarchical decision-making unit needs to have efficient data analysis and decision-making capabilities. This can be achieved by adopting advanced control algorithms and decision-making mechanisms. For example, the hierarchical decision-making unit can select an obstacle avoidance strategy based on the movement speed and intrusion direction of the obstacle. When the intrusion speed is lower than the threshold, it triggers a path replanning algorithm to generate an obstacle avoidance path; when the intrusion speed is higher than the threshold, it activates an electromagnetic brake to lock each joint and sends a hierarchical warning signal to the operator through a tactile feedback unit.

[0092] The self-learning optimization unit needs to have efficient learning and optimization capabilities. This can be achieved by adopting reinforcement learning algorithms. For example, the self-learning optimization unit can train a reinforcement learning model based on historical obstacle avoidance data and dynamically adjust the path planning strategy to improve the obstacle avoidance efficiency. The reinforcement learning model can be trained with a large amount of training data to improve its accuracy and robustness.

[0093] The design of the obstacle avoidance module also needs to consider its real-time performance and reliability. This can be achieved by adopting high-performance processors and sensors. For example, the obstacle avoidance module can be equipped with a real-time operating system to ensure the real-time processing of data and the rapid response of decision-making. In addition, the obstacle avoidance module can also be equipped with a redundant design to improve the reliability of the system.

[0094] The task planning module includes a power grid topology analysis unit, a multi-objective optimization unit, a collaborative control unit, and a dynamic priority adjustment unit. The power grid topology analysis unit imports the 3D point cloud data of the distribution network line and identifies the spatial coordinates and defect levels of insulators, wire joints, and lightning arresters. The multi-objective optimization unit combines the kinematic constraints of the robotic arm, the flight time of the UAV, and the priority of equipment defects to generate an inspection sequence that takes into account both safety and efficiency. The collaborative control unit decomposes the inspection sequence into the flight trajectory of the UAV and the action instructions of the robotic arm, and uses a Kalman filter to correct the vibration amplitude of the end of the robotic arm in real time. The dynamic priority adjustment unit reallocates the operation sequence and updates the flight path when a sudden equipment failure is detected.

[0095] The design of the task planning module needs to consider multiple aspects. First, its power grid topology analysis unit needs to have efficient data import and analysis capabilities. This can be achieved by adopting advanced data processing algorithms and sensor technologies. For example, the power grid topology analysis unit can import the 3D point cloud data of the distribution network line and identify the spatial coordinates and defect levels of insulators, wire joints, and lightning arresters through data processing algorithms. In addition, the power grid topology analysis unit can be equipped with sensors such as lidar and vision sensors to obtain more accurate environmental information.

[0096] The multi-objective optimization unit needs to have efficient data analysis and optimization capabilities. This can be achieved by adopting advanced optimization algorithms and control strategies. For example, the multi-objective optimization unit can combine the kinematic constraints of the robotic arm, the flight time of the UAV, and the priority of equipment defects to generate an inspection sequence that takes into account both safety and efficiency. Optimization algorithms such as genetic algorithms and particle swarm optimization algorithms can be used to improve the efficiency and accuracy of optimization.

[0097] The collaborative control unit needs to have efficient data decomposition and control capabilities. This can be achieved by adopting advanced control algorithms and collaborative mechanisms. For example, the collaborative control unit can decompose the inspection sequence into the flight trajectory of the UAV and the action instructions of the robotic arm, and use a Kalman filter to correct the vibration amplitude of the end of the robotic arm in real time. In addition, the collaborative control unit can be equipped with a real-time operating system to ensure the real-time processing of data and the rapid response of control.

[0098] The dynamic priority adjustment unit needs to have efficient data analysis and adjustment capabilities. This can be achieved by adopting advanced control algorithms and decision-making mechanisms. For example, the dynamic priority adjustment unit can reallocate the operation sequence and update the flight path when a sudden equipment failure is detected. In addition, the dynamic priority adjustment unit can be equipped with a real-time monitoring mechanism to ensure the safe operation of the system.

[0099] The design of the grasping fixture takes into account high precision and safety. The multi-modal tactile perception unit detects the shape, hardness, and surface material distribution of the target object through a piezoresistive sensor array and a capacitive sensor. The fuzzy adaptive clamping unit dynamically adjusts the opening and closing angle of the jaws and the clamping force according to the tactile data, making the contact pressure distribution uniform and the peak pressure less than the preset threshold. The dual-mode insulation monitoring unit works in parallel through an electric field sensor and a leakage current detection circuit, triggering a safety lock when the detected potential difference or leakage current exceeds the threshold. The self-cleaning mechanism removes foreign objects on the contact surface through high-pressure air flow when the jaws are closed, ensuring the effective isolation of the insulation protection layer.

[0100] The design of the grasping fixture needs to consider multiple aspects. First of all, its multi-modal tactile perception unit needs to have efficient data acquisition and processing capabilities. This can be achieved by adopting multiple sensor technologies. For example, a piezoresistive sensor array can be used to detect the shape and hardness of the target object, while a capacitive sensor can be used to detect the surface material distribution of the target object. Through the fusion processing of the data of these sensors, a comprehensive perception of the target object can be realized.

[0101] The fuzzy adaptive clamping unit needs to have efficient data analysis and clamping capabilities. This can be achieved by adopting advanced control algorithms and actuator technologies. For example, the fuzzy adaptive clamping unit can dynamically adjust the opening and closing angle of the jaws and the clamping force according to the tactile data, making the contact pressure distribution uniform and the peak pressure less than the preset threshold. In addition, the fuzzy adaptive clamping unit can also be equipped with a micro servo motor to achieve precise clamping control.

[0102] The dual-mode insulation monitoring unit needs to have efficient data monitoring and safety protection capabilities. This can be achieved by adopting an electric field sensor and a leakage current detection circuit. For example, an electric field sensor can be used to detect the potential difference, while a leakage current detection circuit can be used to detect the leakage current. When the detected potential difference or leakage current exceeds the threshold, the dual-mode insulation monitoring unit can trigger a safety lock to ensure the safe operation of the system.

[0103] The self-cleaning mechanism needs to have efficient data processing and cleaning capabilities. This can be achieved by adopting high-pressure air flow technology and control algorithms. For example, the self-cleaning mechanism can remove foreign objects on the contact surface through high-pressure air flow when the jaws are closed, ensuring the effective isolation of the insulation protection layer. In addition, the self-cleaning mechanism can also be equipped with a real-time monitoring mechanism to ensure the cleaning effect.

[0104] The design of the spraying device takes into account efficient spraying and quality control. The ultrasonic viscosity detection unit monitors the rheological properties of the paint in real time through a sound velocity sensor, generating a viscosity-temperature correlation curve. The dynamic atomizing nozzle adjusts the nozzle aperture and atomizing air pressure through a PID algorithm according to the flight speed of the drone, the curvature of the target surface, and the viscosity-temperature curve, so that

[0105] The paint deposition uniformity reaches over 90%. The coating thickness feedback unit uses a laser interferometer to detect the spraying thickness in real time and triggers the supplementary spraying program when the thickness deviation exceeds the threshold. When the self-repairing nozzle detects nozzle blockage, it clears the sediment through reverse pulse airflow and restores the atomization performance.

[0106] The design of the spraying device needs to consider multiple aspects. First, its ultrasonic viscosity detection unit needs to have efficient data acquisition and processing capabilities. This can be achieved by adopting advanced sensor technology and signal processing algorithms. For example, the ultrasonic viscosity detection unit can monitor the rheological properties of the paint in real time through a sound velocity sensor and generate a viscosity-temperature correlation curve. In addition, the ultrasonic viscosity detection unit can also be equipped with data processing algorithms to improve the accuracy and reliability of the data.

[0107] The dynamic atomizing nozzle needs to have efficient data analysis and spraying capabilities. This can be achieved by adopting advanced control algorithms and driver technology. For example, the dynamic atomizing nozzle can adjust the nozzle aperture and atomizing air pressure through the PID algorithm according to the UAV flight speed, the curvature of the target surface, and the viscosity-temperature curve, so that the paint deposition uniformity reaches over 90%. In addition, the dynamic atomizing nozzle can also be equipped with a micro servo motor to achieve precise spraying control.

[0108] The coating thickness feedback unit needs to have efficient data monitoring and feedback capabilities. This can be achieved by adopting a laser interferometer and control algorithms. For example, the coating thickness feedback unit can detect the spraying thickness in real time through a laser interferometer and trigger the supplementary spraying program when the thickness deviation exceeds the threshold. In addition, the coating thickness feedback unit can also be equipped with a real-time monitoring mechanism to ensure the spraying quality.

[0109] The self-repairing nozzle needs to have efficient data processing and repair capabilities. This can be achieved by adopting reverse pulse airflow technology and control algorithms. For example, when the self-repairing nozzle detects nozzle blockage, it can clear the sediment through reverse pulse airflow and restore the atomization performance. In addition, the self-repairing nozzle can also be equipped with a real-time monitoring mechanism to ensure the repair effect.

[0110] The design of the laser processing unit takes into account efficient cleaning and safety. The spectral material identification module analyzes the composition of the target foreign object through reflected spectroscopy and matches the laser wavelength-power combination table in the preset database. The dynamic focusing module adjusts the beam divergence angle in real time according to the distance of the foreign object, so that the energy density error is controlled within the preset range. When the safety interlock unit detects that a person enters the operation radius of 2m, it reduces the laser power to the safety level and generates an electronic fence warning signal to be linked and locked with the physical shielding cover. The energy distribution optimization module optimizes the laser scanning path based on the foreign object material and thickness distribution through the genetic algorithm to improve the cleaning efficiency.

[0111] The design of the laser processing unit needs to consider multiple aspects. First of all, its spectral material recognition module needs to have efficient data acquisition and processing capabilities. This can be achieved by adopting advanced sensor technologies and signal processing algorithms. For example, the spectral material recognition module can analyze the composition of the target foreign object through reflection spectroscopy and match the laser wavelength-power combination table in the preset database. In addition, the spectral material recognition module can also be equipped with data processing algorithms to improve the accuracy and reliability of the data.

[0112] The dynamic focusing module needs to have efficient data analysis and focusing capabilities. This can be achieved by adopting advanced control algorithms and driver technologies. For example, the dynamic focusing module can adjust the beam divergence angle in real time according to the distance of the foreign object, so that the energy density error is controlled within the preset range. In addition, the dynamic focusing module can also be equipped with a micro servo motor to achieve precise focusing control.

[0113] The safety interlock unit needs to have efficient data monitoring and safety protection capabilities. This can be achieved by adopting advanced sensor technologies and control algorithms. For example, when the safety interlock unit detects that a person enters the operation radius of 2m, it can reduce the laser power to the safety level and generate an electronic fence warning signal to be linked and locked with the physical shielding cover. In addition, the safety interlock unit can also be equipped with a real-time monitoring mechanism to ensure the safe operation of the system.

[0114] The energy distribution optimization module needs to have efficient data processing and optimization capabilities. This can be achieved by adopting genetic algorithms and control strategies. For example, based on the material and thickness distribution of the foreign object, the energy distribution optimization module can optimize the laser scanning path through genetic algorithms to improve the cleaning efficiency. In addition, the energy distribution optimization module can also be equipped with a real-time monitoring mechanism to ensure the optimization effect.

[0115] Through the above detailed technical feature explanations and subordination expansions, the multi-functional robotic arm system for distribution network UAV inspection of the present invention can achieve efficient and accurate inspection operations, and at the same time has good reliability and safety.

[0116] It is known by common technical knowledge that the present invention can be implemented by other embodiments that do not depart from its spiritual essence or essential features. Therefore, the above-disclosed embodiments are illustrative in all aspects and are not the only ones. All changes within the scope of the present invention or within the scope equivalent to the present invention are encompassed by the present invention.

Claims

1. A multi-functional robotic arm system for distribution network UAV inspection, characterized in that, Comprising: A drone platform for providing flight control and load-bearing functions; A robotic arm module detachably connected below the drone platform, including a multi-degree-of-freedom joint structure and a drive unit, the robotic arm module being used to perform grasping, spraying, and laser operations; A bionic control system, including a motion capture unit and an instruction conversion unit, the motion capture unit real-time collecting the motion trajectory data of the operator's arm, and converting the motion trajectory data into robotic arm module action instructions through the instruction conversion unit to achieve synchronous motion control of the robotic arm module and the operator's arm; Replaceable operation components, including a variety of operation components, connected to the end of the robotic arm module through a standardized interface.

2. The system according to claim 1, wherein The bionic control system further includes: An electromyogram signal analysis unit for collecting the electromyogram signals of the operator's arm through a bioelectric sensor of a wearable device, and performing time-domain noise suppression processing on the electromyogram signals to generate a denoised electromyogram waveform; A motion trajectory reconstruction unit for inputting the denoised electromyogram waveform into a pre-trained neural network model to output a three-dimensional motion trajectory vector of the operator's arm, the neural network model being trained based on historical operation data to eliminate signal distortion caused by environmental electromagnetic interference; A joint dynamic mapping unit for generating rotation angle, angular velocity, and torque instructions for each joint according to the inverse solution mapping relationship between the three-dimensional motion trajectory vector and the robotic arm joint kinematic model, and transmitting them to the main control unit of the drone platform through a wireless communication module; A delay compensation module for predicting the instruction transmission delay based on the wireless communication link state, and preloading action instructions in the robotic arm joint controller through a feedforward control algorithm to eliminate operation lag.

3. The system according to claim 1, wherein The joint structure of the robotic arm module includes: A harmonic reduction rotating joint for providing continuous rotational motion in the horizontal plane through a strain wave gear, with a multi-axis torque sensor integrated at the end of the harmonic reducer to real-time monitor the rotational load torque and feedback it to the dynamic drive unit; A planetary roller pitching joint for achieving 0°-90° pitching swing through a roller screw pair, with a vibration damping layer provided at the end of the screw to suppress positioning errors caused by high-frequency vibrations; A magnetic levitation telescopic joint for driving a telescopic arm to move along the axis through a linear motor, with a non-contact Hall displacement sensor integrated at the end of the telescopic arm to real-time detect the displacement and feedback it to the dynamic drive unit; A temperature-torque coupling monitoring unit for real-time collecting the associated data of the motor temperature and output torque of each joint, and triggering a frequency reduction protection mechanism and generating an alarm signal when detecting abnormal temperature or torque mutation.

4. The system according to claim 1, wherein The modular interface of the replaceable operation components includes: A multi-modal positioning mechanism for obtaining the three-dimensional attitude data of the operation component during component installation by the collaborative work of a lidar and a near-field communication chip; An active compensation unit for driving a micro servo motor to adjust the position of the interface pin based on the three-dimensional attitude data, making the installation angle error of the component less than ±0.5°, and triggering an electrical connection self-check program to verify the contact reliability. A redundant locking unit, which is used to trigger a secondary locking action through a shape memory alloy after magnetic attraction positioning and mechanical buckling are completed, so as to ensure anti-vibration stability during high-altitude operations.

5. The system according to claim 1, wherein The driving unit further includes: A load prediction module, which is used to analyze historical load data through a long short-term memory network and predict the torque requirements of each joint within the next 5 seconds; A dynamic allocation module, which is used to allocate motor power in advance based on the prediction results and optimize the start-stop timing of the energy recovery circuit; an overload protection circuit, which is used to cut off the power supply of the motor and activate the emergency braking device of the robotic arm when it detects that the instantaneous current exceeds the safety threshold; A bidirectional energy management unit, which is used to preferentially use battery power during the acceleration stage of the robotic arm and store the kinetic energy recovered into electrical energy in the drone battery during the deceleration stage to achieve dynamic energy balance.

6. The system according to claim 1, wherein The obstacle avoidance module includes: A multi-source perception fusion unit, which is used to construct a dynamic environment model within a range of 0.5m - 3m around the robotic arm through a millimeter-wave radar, an infrared thermal imager, and a depth camera; A hierarchical decision-making unit, which is used to select one of the following operations according to the movement speed and intrusion direction of the obstacle: When the intrusion speed is lower than the threshold, trigger a path replanning algorithm to generate an obstacle avoidance path; When the intrusion speed is higher than the threshold, activate the electromagnetic brake to lock each joint and send a hierarchical warning signal to the operator through the tactile feedback unit; A self-learning optimization unit, which is used to train a reinforcement learning model based on historical obstacle avoidance data and dynamically adjust the path planning strategy to improve the obstacle avoidance efficiency.

7. The system according to claim 1, wherein The task planning module further includes: A power grid topology analysis unit, which is used to import the three-dimensional point cloud data of the distribution network line, identify the spatial coordinates and defect levels of insulators, wire joints, and lightning arresters; A multi-objective optimization unit, which is used to generate an inspection sequence that takes into account both safety and efficiency by combining the kinematic constraints of the robotic arm, the flight time of the drone, and the priority of equipment defects; A cooperative control unit, which is used to decompose the inspection sequence into the flight trajectory of the drone and the action instructions of the robotic arm, and continuously correct the vibration amplitude of the end of the robotic arm through a Kalman filter; A dynamic priority adjustment unit, which is used to reallocate the operation sequence and update the flight path when a sudden equipment failure is detected.

8. The system according to claim 1, wherein The grasping fixture includes: A multi-modal tactile perception unit, which is used to detect the shape, hardness, and surface material distribution of the target object through a piezoresistive sensor array and a capacitive sensor; A fuzzy adaptive clamping unit, which is used to dynamically adjust the opening and closing angle of the jaws and the clamping force according to the tactile data, so that the contact pressure distribution is uniform and the peak pressure is less than the preset threshold; A dual-mode insulation monitoring unit, which is used to work in parallel through an electric field sensor and a leakage current detection circuit, and trigger a safety lock when it detects that the potential difference exceeds the voltage threshold or the leakage current exceeds the current threshold; A self-cleaning mechanism, which is used to remove foreign objects on the contact surface through high-pressure air flow when the jaws are closed to ensure the effective isolation of the insulation protection layer.

9. The system according to claim 1, characterized in that The spraying device includes: An ultrasonic viscosity detection unit, which is used to continuously monitor the rheological properties of the paint through a sound velocity sensor and generate a viscosity-temperature correlation curve; A dynamic atomizing nozzle, which is used to adjust the nozzle aperture and atomizing air pressure according to the flight speed of the drone, the curvature of the target surface and the viscosity-temperature curve through a PID algorithm, so as to make the coating deposition uniformity reach more than 90%; A coating thickness feedback unit, which is used to detect the spraying thickness in real time through a laser interferometer and trigger a supplementary spraying program when the thickness deviation exceeds the thickness threshold; A self-repairing nozzle head, which is used to clear the deposits and restore the atomizing performance through reverse pulse air flow when a nozzle blockage is detected.

10. The system according to claim 1, wherein The laser processing unit includes: A spectral material identification module, which is used to analyze the composition of the target foreign object through reflection spectroscopy and match the laser wavelength-power combination table in the preset database; A dynamic focusing module, which is used to adjust the beam divergence angle in real time according to the distance of the foreign object, so as to control the energy density error within a preset range; A safety interlock unit, which is used to reduce the laser power to a safe level when it detects that a person enters a working radius of 2 m, and generate an electronic fence warning signal to be linked and locked with a physical shielding cover; An energy distribution optimization module, which is used to optimize the laser scanning path based on the foreign object material and thickness distribution through a genetic algorithm to improve the cleaning efficiency.

Images