Motion track planning and obstacle avoidance method for wheeled two-arm humanoid robot

Abstract

The invention relates to a wheel type double-arm humanoid robot motion track planning and obstacle avoidance method. The method comprises the steps that a robot system is started for self-inspection; the first sensor obtains environment information and constructs a real-time three-dimensional point cloud obstacle information graph; the second sensor obtains chassis state information and mechanical arm state information; a target task is set, and a moving path of a chassis and a moving path of a mechanical arm are planned based on the state information of the target task; and the system controls the chassis to move according to the planned moving path through the cooperative control model, and the mechanical arm moves according to the planned moving track. Through construction and real-time updating of the real-time three-dimensional point cloud obstacle information graph, the robot can effectively recognize and avoid obstacles, the risk of collision with the surrounding environment is reduced, meanwhile, the robot can adapt to the dynamically changing environment, path planning is rapidly updated through the D * Lite algorithm, and the path planning efficiency is improved. And the path can be adjusted in time when the obstacle moves or a new obstacle appears.

Description

Technical Field

[0001] The invention belongs to the technical field of robots, and in particular relates to a motion trajectory planning and obstacle avoidance method for a wheeled dual-arm humanoid robot. Background Art

[0002] With the rapid development and continuous breakthroughs in robotics technology, the application scope of humanoid robots that combine wheeled mobility and dual-arm operation capabilities in many fields is rapidly expanding. However, in actual applications, when faced with complex terrain, robots may not be able to accurately judge the terrain features, resulting in unstable driving or even failures; when encountering dynamic obstacles, their reaction speed and decision-making ability are often not enough to make quick adjustments, thereby increasing the risk of collision; and when operating in narrow spaces, the robot's movement is restricted, and traditional planning algorithms often find it difficult to find the optimal path, resulting in low efficiency.

[0003] Therefore, a motion trajectory planning and obstacle avoidance method for a wheeled dual-arm humanoid robot is designed so that the robot can quickly perceive changes in the surrounding environment in a rapidly changing environment and immediately adjust its own motion strategy to ensure that it safely reaches the target mission location. Summary of the invention

[0004] The object of the present invention is to provide a motion trajectory planning and obstacle avoidance method for a wheeled dual-arm humanoid robot to solve the problems raised in the above background technology.

[0005] To achieve the above object, the present invention provides the following technical solution: a motion trajectory planning and obstacle avoidance method for a wheeled dual-arm humanoid robot, the steps comprising:

[0006] S01 starts the robot system to perform self-check;

[0007] S02 The first sensor acquires environmental information and constructs a real-time three-dimensional point cloud obstacle information map;

[0008] S03 The second sensor obtains chassis status information and robot arm status information;

[0009] S04 sets a target task, and plans a moving path of the chassis and a motion path of the robotic arm based on the state information of the target task;

[0010] The S05 system controls the chassis to move according to the planned moving path and the robotic arm to move according to the planned motion trajectory through the collaborative control model;

[0011] The second sensor S06 detects the environmental data and robot status in real time, updates the second planned working path through the D*Lite algorithm based on the real-time environmental data and additional constraints, and generates a real-time moving path for the chassis;

[0012] S07 arrives at the target task position, the chassis is adjusted based on the object status information, and cooperates with the robotic arm to grab the object from the initial position.

[0013] Preferably, the first sensor acquires environmental information and constructs a real-time three-dimensional point cloud obstacle information map, wherein:

[0014] The first sensor is a laser radar;

[0015] LiDAR emits a laser beam and analyzes the reflected signal to draw a three-dimensional point cloud image of the surrounding environment, with each point representing a specific location in space;

[0016] Use point cloud segmentation algorithms to identify and track obstacles, extract their location information from the point cloud, mark them in the navigation system, and generate a three-dimensional point cloud obstacle information map.

[0017] Preferably, the second sensor acquires chassis status information and mechanical arm status information including:

[0018] The second sensor is an accelerometer and a gyroscope;

[0019] The chassis status information specifically includes the chassis position, speed and posture information;

[0020] The state information of the robotic arm specifically includes the positions of both arms and the motion data of both arms.

[0021] Preferably, the target task is set by planning the moving path of the chassis based on the state information of the target task, wherein the state information of the target task includes position information and space information.

[0022] Preferably, the setting of the target task and planning the moving path of the chassis and the motion path of the robotic arm based on the state information of the target task include:

[0023] S401 plans a target path for the chassis to move based on the position information of the target task, and uses a path planning algorithm to obtain a first planned working path;

[0024] S402 plans the motion trajectory of the robot arm based on the spatial information of the target task, and calculates the maximum contour range of the robot arm during the motion process through inverse kinematics;

[0025] S403 inputs the obtained robot arm motion profile into the environment model to generate an extended obstacle;

[0026] S404: Optimize the first planned working path using a path planning algorithm based on the generated extended obstacles to obtain a second planned working path.

[0027] Preferably, the path planning algorithm is an A* algorithm.

[0028] Preferably, the system controls the chassis to move according to the planned moving path and the robot arm to move according to the planned motion trajectory through the collaborative control model.

[0029] The synchronous motion information of the chassis includes speed, direction and motion trajectory

[0030] The synchronous motion information of the robot arm includes extension, rotation and any possible swing of the arms.

[0031] Preferably, the second sensor detects environmental data and robot status in real time, and based on the real-time environmental data and additional constraints, updates the second planned working path through the D*Lite algorithm to generate a real-time moving path of the chassis, wherein the additional constraints are extended obstacles calculated in real time.

[0032] Preferably, the robot also includes a memory, a processor, and a path planning and obstacle avoidance program of the robot stored in the memory and executable on the processor, and the path planning and obstacle avoidance program of the robot is configured to implement the steps of the obstacle avoidance method of the industrial robot as described in any one of claims 1 to 8.

[0033] Compared with the prior art, the present invention has the following beneficial effects:

[0034] The present invention enables the robot to effectively identify and avoid obstacles and reduce the risk of collision with the surrounding environment by constructing and updating real-time three-dimensional point cloud obstacle information maps. At the same time, the robot can adapt to dynamically changing environments and quickly update path planning through the D*Lite algorithm to ensure that the path can be adjusted in time when obstacles move or new obstacles appear.

[0035] The present invention realizes independent control of the robot chassis and the mechanical arm through a collaborative control model, and realizes collaborative work through a communication protocol, thereby improving the efficiency of task execution. At the same time, by using the chassis and mechanical arm status information obtained by sensors, the robot can accurately control its movement and ensure that the mechanical arm end effector can accurately reach the position of the target object;

[0036] The present invention uses data fusion of multiple sensors and systems, such as laser radar, camera, infrared sensor, etc., so that the robot can obtain a more complete and realistic three-dimensional environment representation, thereby enhancing the robustness of the system;

[0037] The present invention optimizes the working path through the path planning algorithm, so that the robot can reach the target position in the shortest path, saving time and energy. At the same time, the actuator parameters are adjusted according to the state information of the target object to ensure that the object can be moved safely and accurately from the initial position to the target position.

[0038] The robot design in the present invention includes a processor, a communication bus, a user interface and a network interface, so that it can be compatible with a variety of external devices and network environments, and is easy to expand and integrate. The memory setting enables the robot to store and execute programs, and at the same time store parameters for use by the industrial robot, thereby improving the efficiency and stability of data processing. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 Flow chart of a motion trajectory planning and obstacle avoidance method for a wheeled dual-arm humanoid robot in an embodiment. DETAILED DESCRIPTION

[0040] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0041] Example 1

[0042] The obstacle avoidance method in this embodiment is applied to an industrial robot, wherein the industrial robot can be a multi-joint manipulator or a humanoid robot, wherein the humanoid robot has a chassis and a dual-arm manipulator, and the chassis moves by executing a navigation module to drive the entire robot to move. Specifically, the robot's motion trajectory planning and obstacle avoidance method in this embodiment includes:

[0043] S01 starts the robot system to perform self-check;

[0044] S02 The first sensor acquires environmental information and constructs a real-time three-dimensional point cloud obstacle information map;

[0045] S03 The second sensor obtains chassis status information and robot arm status information;

[0046] S04 sets a target task, and plans a moving path of the chassis and a motion path of the robotic arm based on the state information of the target task;

[0047] The S05 system controls the chassis to move according to the planned moving path and the robotic arm to move according to the planned motion trajectory through the collaborative control model;

[0048] The second sensor S06 detects the environmental data and robot status in real time, updates the second planning path through the D*Lite algorithm based on the real-time environmental data and additional constraints, and generates a real-time moving path for the chassis;

[0049] S07 arrives at the target task position, the chassis is adjusted based on the object status information, and cooperates with the robotic arm to grab the object from the initial position.

[0050] In one embodiment, step S01 starts the robot system to perform self-check, which specifically includes:

[0051] During the system startup phase, it is usually necessary to perform self-tests on the hardware and software in each module. This includes power supply checks, whether the moving joints of the robot arm and chassis are flexible, without jamming or damage, and calibration of each sensor installed in the system to ensure that the data acquired is accurate, as well as whether the navigation system of the chassis and the motion system of the robot arm are operating normally, etc.

[0052] In one embodiment, S02, the first sensor acquires environmental information and constructs a real-time three-dimensional point cloud obstacle information map; specifically, it includes:

[0053] The first sensor is a laser radar;

[0054] LiDAR emits a laser beam and analyzes the reflected signal to draw a three-dimensional point cloud image of the surrounding environment, with each point representing a specific location in space;

[0055] Use point cloud segmentation algorithms to identify and track obstacles, extract their location information from the point cloud, mark them in the navigation system, and generate a three-dimensional point cloud obstacle information map.

[0056] It can be understood that after starting, the robot in this embodiment starts from the initial position, navigates to the target position, drives the arms to grab objects, and moves in a point-to-point predicted trajectory. In the predicted trajectory, people and objects that appear within the robot's obstacle avoidance range are collectively referred to as obstacles. The system needs to obtain data based on the obstacles and re-plan the working path.

[0057] Among them, the first sensor can be any sensor that can be used to collect surrounding environment data, such as laser radar, camera, infrared sensor, etc. In this embodiment, the first sensor is limited to laser radar. By emitting a series of rapidly rotating laser beams to the surrounding environment, the laser radar receiver captures the signal reflected back when the laser beam encounters an obstacle, and by analyzing the intensity, time and direction of these reflected signals, the three-dimensional coordinates of each reflection point relative to the sensor are accurately calculated, thereby drawing a high-precision three-dimensional point cloud image of the surrounding environment. In this image, each point represents a specific position in space.

[0058] Then use the point cloud segmentation algorithm to identify and track obstacles from the known point cloud data. It can be understood that obstacles are divided into animals and still objects. Still objects remain motionless in the image and can be used as basic data for preliminary path planning. Animals need to be marked in the image and keep tracking obstacles after moving to achieve effective obstacle avoidance. The point cloud segmentation algorithm can divide the point cloud data into different clusters or objects based on information such as the geometric characteristics of the points, spatial distribution, and reflection intensity of the laser beam. Each cluster represents a potential obstacle. By continuously tracking the changes in these clusters, the position, shape, and movement speed of the obstacle can be extracted in real time. The obstacle information is input into the navigation system for marking and positioning to generate a three-dimensional point cloud obstacle information map, which is used to provide data support and reference for the robot's path planning.

[0059] In addition, the construction of a three-dimensional point cloud obstacle information map can be realized through a stereo vision system. Two or more cameras are used to simultaneously shoot the same scene from different angles, and the three-dimensional coordinates of each point in the scene are calculated using the principle of parallax. Based on the installation and use of lidar, a three-dimensional point cloud obstacle information map can also be constructed in coordination with the shooting data of the structured light system and the camera. The lidar in the structured light system projects a beam of light of a known pattern to the surrounding environment, and then uses the deformation of the pattern captured by the camera to restore the three-dimensional shape of the scene by calculating the difference between the deformed pattern and the original pattern. It is understandable that after the data of the above-mentioned multiple sensors and systems are fused, for example, the high-precision distance measurement of the lidar and the rich color information of the camera can be integrated to obtain a more complete and realistic three-dimensional environment representation. By integrating their respective advantages, the accuracy and robustness of the three-dimensional point cloud obstacle information map can be improved.

[0060] In one embodiment, in step S03, the second sensor obtains chassis status information and robot arm status information; specifically:

[0061] The chassis status information specifically includes the chassis position, speed and attitude information;

[0062] The status information of the robot arm specifically includes the position of both arms and the motion data of both arms.

[0063] The second sensor is any sensor that can measure the angular velocity and acceleration of an object, such as an accelerometer and a gyroscope. The gyroscope is mainly installed in the chassis. The change in the chassis's posture can be obtained by integrating the angular velocity data. Combined with the initial posture information, the current posture of the chassis can be determined. Combined with the data obtained by the accelerometer, the linear velocity and position information of the chassis can be indirectly calculated. The calculation steps can be specifically referred to as follows: In a static or uniform linear motion state, the value measured by the accelerometer is the component of the gravitational acceleration g, while in a dynamic state, the value measured by the accelerometer is the synthesis of the gravitational acceleration and the chassis motion acceleration. Therefore, in order to calculate the chassis's linear velocity and position information from the acceleration data, it is necessary to first remove the influence of the gravitational acceleration, obtain the chassis's motion acceleration relative to the ground, and then integrate the motion acceleration to obtain the chassis's linear velocity; and then integrate the linear velocity to obtain the chassis's position information.

[0064] The formulas involved are:

[0065] a=[ax,ay,az](unit: m / s2), then v˙=a.

[0066] Among them, a is the motion acceleration vector of the chassis, and v is the linear velocity vector.

[0067] v=[vx,vy,vz](unit: m / s), then p˙=v

[0068] Where p is the position vector of the chassis.

[0069] Secondly, the accelerometer or gyroscope can be used to obtain the state information of the robot arm, which specifically includes the initial position of the two arms and the motion data of the two arms. The system obtains the initial motion data of the two arms from the database, including but not limited to the joint angle, speed and strength information of the robot arm. After the robot arm is started, the sensors installed on each joint of the robot arm can be used to measure the angular velocity of the joint, and the rotation angle of the joint and the overall posture change of the robot arm can be obtained by integrating the angular velocity data. In addition, a torque sensor can be loaded to measure the torque applied to the joint. After analyzing the measurement data, the load and stress applied to the robot arm during the movement can be understood, so as to achieve more precise control and adjustment.

[0070] In one embodiment, step S04: setting a target task, and planning a moving path of the chassis and a motion path of the robot arm based on the state information of the target task; specifically includes:

[0071] S401 plans a target path for the chassis to move based on the position information of the target task, and uses a path planning algorithm to obtain a first planned working path;

[0072] S402 plans the motion trajectory of the robot arm based on the spatial information of the target task, and calculates the maximum contour range of the robot arm during the motion process through inverse kinematics;

[0073] S403 inputs the obtained robot arm motion profile into the environment model to generate an extended obstacle;

[0074] S404: Optimize the first planned working path using a path planning algorithm based on the generated extended obstacles to obtain a second planned working path.

[0075] First, clarify the specific task that the robot needs to complete, which in this embodiment is to carry objects. Pre-set the target task position that the robot needs to move, that is, the position of the object to be carried. Based on the construction of the three-dimensional point cloud obstacle information map, obtain the relative position of the target task position relative to the current position of the robot. Calculate the optimal movement path of the chassis based on the map information, obstacle position and target position, that is, the first planned working path. Secondly, based on the spatial information of the target task, determine the optimal motion trajectory of the robot arm when completing the target task. Pre-acquire the spatial information of the target task, including the size, shape and required grasping operation of the object to be carried, etc. Through inverse kinematics calculation, determine the joint angle and speed of the robot arm when performing the task, so as to plan the motion trajectory of the robot arm. During this period, calculate the maximum contour range of the robot arm during the movement, evaluate the space that the robot arm may occupy during the movement, and avoid collision with obstacles.

[0076] For the n-degree-of-freedom (DOF) serial robot in this embodiment, it is necessary to calculate the joint angles of the robot based on the Denavit-Hartenberg (DH) parameters to determine the joint angle data of the robot when the joint angle of the robot reaches a specific end position, that is, reaches the maximum contour range, and then calculate the maximum contour range of the robot.

[0077] Understandably, considering the space that the robot may occupy during the movement, the calculated maximum contour range of the robot is used as an extended obstacle, and the known three-dimensional point cloud obstacle information map is input, which includes all possible positions of the robot when performing the task, and the three-dimensional point cloud obstacle information map is updated to include the robot as an "extended obstacle"; based on the update of the three-dimensional point cloud obstacle information map, the path planning algorithm is used again to optimize the first planned working path to generate the second planned working path. The second planned working path reflects the actual movement requirements of the robot when performing the task. The optimized path ensures that the chassis will not collide with the extended obstacle during the movement, and maintains safety when executing the planned path.

[0078] It should be noted that the path planning algorithm of the present invention mainly relies on the A* algorithm to find the shortest path from the initial position to the target task position on the three-dimensional point cloud obstacle information map through heuristic search.

[0079] In one embodiment, the S05 system controls the chassis to move according to the planned moving path and the robot arm to move according to the planned motion trajectory through the collaborative control model; specifically, it includes:

[0080] This embodiment uses a collaborative control model to separate the controllers of the chassis and the robotic arm, and realizes collaborative control through communication protocols such as the CAN bus, including sending motion instructions to the chassis and the robotic arm to execute a pre-planned movement path, and during the movement, the chassis and the robotic arm make real-time adjustments based on sensor feedback to cope with environmental changes.

[0081] In collaborative control, the robot moves and exercises according to the second planned working path, and at the same time integrates the motion information of the chassis and the robotic arm by developing a sensor fusion algorithm to improve the overall accuracy and reliability of the system. Among them, the synchronous motion information of the chassis includes speed, direction and motion trajectory, and the synchronous motion information of the robotic arm includes the extension, rotation and any possible swing of the arms.

[0082] Among them, the CAN bus communication protocol can specifically plan the movement paths of the chassis and robotic arms according to service requirements, and send movement instructions to the controllers of the chassis and robotic arms through the CAN bus. During the movement, the chassis and robotic arms detect environmental changes in real time through their respective sensors, and feed back the detection information to the control system through the CAN bus. The control system makes real-time adjustments based on the feedback information to ensure that the chassis and robotic arms can move accurately along the predetermined path.

[0083] In one embodiment, in step S06, the second sensor detects the environmental data and the robot state in real time, and based on the real-time environmental data and the additional constraints, updates the second planned path through the D*Lite algorithm to generate a real-time moving path for the chassis, specifically including:

[0084] The robot in this embodiment is in a factory or outdoor environment where a surrounding environment map is not given in advance. After startup, the robot starts from the initial position, navigates to the target position, drives the arms to grab the object, and moves in a point-to-point predicted trajectory. During the movement of the robot executing the second planned path, the second sensor continuously updates the robot's posture data, and the first sensor continuously detects environmental changes, such as the movement of obstacles or new obstacles. Based on the obstacle information and the expansion of the obstacle avoidance range, additional points are determined, and the second planned path is replanned according to the additional points and the target task points.

[0085] Among them, the update of the second planned path relies on the D*Lite algorithm, which improves the efficiency of path planning by incrementally updating path information. It can quickly update path planning according to real-time road conditions, so that the robot can quickly adjust the path when facing obstacles or environmental changes to ensure safe arrival at the target task point. Compared with the A algorithm, which needs to replan the path from scratch in a dynamic environment, the D Lite algorithm can quickly adjust the path through incremental updates, and has higher path planning efficiency in a dynamic environment.

[0086] The D Lite algorithm recalculates a more optimal moving path based on environmental changes and new obstacles. The chassis control system immediately receives this new path information and responds quickly. Specifically, the control system first evaluates the current motion state of the chassis, and then makes necessary adjustments based on the requirements of the new path, including changing the driving direction, accelerating or decelerating, etc., to ensure that the chassis can move smoothly and quickly along the new path.

[0087] In one embodiment, step S07 reaches the target task position, the chassis is adjusted based on the object state information, and the robot arm is used to grab the object from the initial position; specifically, it includes:

[0088] When the robot moves to the vicinity of the target task position, the robotic arm will make precise posture adjustments based on the position and posture of the chassis, as well as the relative position of the target task position relative to the chassis, to ensure that the end of the robotic arm can accurately reach the initial position of the target object; then, based on the identified target object state information, the system will adjust the actuator at the end of the robotic arm to adapt to the object, adjusting the execution parameters including the size and shape of the object, adjusting the opening size of the actuator, clamping force and other parameters; based on the adjusted execution parameters, the robotic arm will perform a grasping action to move the object from the initial position to the target position.

[0089] Furthermore, after the object is moved to the target position, the system uses a camera or other equipment to take photos or videos of the object at the target position, and compares them with the preset target position to ensure that the posture of the object at the target position is consistent with expectations, without any tilt, inversion or other problems, to ensure that the task has been successfully completed.

[0090] Example 2

[0091] The robot components described in the embodiment of the present invention include: a processor, such as a CPU, a communication bus, a user interface, a network interface and a memory. Among them, the communication bus connects these components and realizes the communication connection of the major components. The user interface design is rich, not only covering the display screen and input unit (such as a keyboard), but also optionally including standard wired interfaces and wireless interfaces. The network interface is also flexible and can selectively integrate standard wired interfaces or wireless interfaces (such as Wi-Fi).

[0092] The memory is configured in the main body of the industrial robot to carry programs that can realize specific functions when executed, and is responsible for storing the parameters required for the operation of the industrial robot. The types of memory are diverse, which can be high-speed RAM memory for performance, stable non-volatile memory (non-volatile memory), such as disk storage, or even a storage unit independent of the processor.

[0093] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0094] The above description is only used to illustrate the technical solution of the present invention rather than to limit it. Other modifications or equivalent substitutions made to the technical solution of the present invention by ordinary technicians in this field should be included in the scope of the claims of the present invention as long as they do not depart from the spirit and scope of the technical solution of the present invention.

Claims

1. A motion trajectory planning and obstacle avoidance method for a wheeled dual-arm humanoid robot, comprising: S01 starts the robot system to perform self-check; S02 The first sensor acquires environmental information and constructs a real-time three-dimensional point cloud obstacle information map; S03 The second sensor obtains chassis status information and robot arm status information; S04 sets a target task, and plans a moving path of the chassis and a motion path of the robotic arm based on the state information of the target task; The S05 system controls the chassis to move according to the planned moving path and the robotic arm to move according to the planned motion trajectory through the collaborative control model; The second sensor S06 detects the environmental data and robot status in real time, updates the second planned working path through the D*Lite algorithm based on the real-time environmental data and additional constraints, and generates a real-time moving path for the chassis; S07 arrives at the target task position, the chassis is adjusted based on the object status information, and cooperates with the robotic arm to grab the object from the initial position.

2. The method for dual-arm motion planning and three-dimensional obstacle avoidance of a wheeled humanoid robot according to claim 1, characterized in that: The first sensor acquires environmental information and constructs a real-time three-dimensional point cloud obstacle information map, wherein: The first sensor is a laser radar; LiDAR emits a laser beam and analyzes the reflected signal to draw a three-dimensional point cloud image of the surrounding environment, with each point representing a specific location in space; Use point cloud segmentation algorithms to identify and track obstacles, extract their location information from the point cloud, mark them in the navigation system, and generate a three-dimensional point cloud obstacle information map.

3. The method for dual-arm motion planning and three-dimensional obstacle avoidance of a wheeled humanoid robot according to claim 1, characterized in that: The second sensor acquires chassis status information and robot arm status information including: The second sensor is an accelerometer and a gyroscope; The chassis status information specifically includes the chassis position, speed and posture information; The state information of the robotic arm specifically includes the positions of both arms and the motion data of both arms.

4. The method for dual-arm motion planning and three-dimensional obstacle avoidance of a wheeled humanoid robot according to claim 1, characterized in that: The target task is set, and a moving path of the chassis is planned based on the state information of the target task, wherein the state information of the target task includes position information and space information.

5. The method for dual-arm motion planning and three-dimensional obstacle avoidance of a wheeled humanoid robot according to claim 4, characterized in that: The setting of the target task and planning the moving path of the chassis and the moving path of the robot arm based on the state information of the target task include: S401 plans a target path for the chassis to move based on the position information of the target task, and uses a path planning algorithm to obtain a first planned working path; S402 plans the motion trajectory of the robot arm based on the spatial information of the target task, and calculates the maximum contour range of the robot arm during the motion process through inverse kinematics; S403 inputs the obtained robot arm motion profile into the environment model to generate an extended obstacle; S404: Optimize the first planned working path using a path planning algorithm based on the generated extended obstacles to obtain a second planned working path.

6. The method for dual-arm motion planning and three-dimensional obstacle avoidance of a wheeled humanoid robot according to claim 5, characterized in that: The path planning algorithm is the A* algorithm.

7. The method for dual-arm motion planning and three-dimensional obstacle avoidance of a wheeled humanoid robot according to claim 1, characterized in that: The system controls the chassis to move according to the planned moving path and the robot arm to move according to the planned motion trajectory through the collaborative control model. The synchronous motion information of the chassis includes speed, direction and motion trajectory The synchronous motion information of the robot arm includes extension, rotation and any possible swing of the arms.

8. The method for dual-arm motion planning and three-dimensional obstacle avoidance of a wheeled humanoid robot according to claim 1, characterized in that: The second sensor detects environmental data and robot status in real time, and based on the real-time environmental data and additional constraints, updates the second planned working path through the D*Lite algorithm to generate a real-time moving path of the chassis, wherein the additional constraints are extended obstacles calculated in real time.

9. The method for dual-arm motion planning and three-dimensional obstacle avoidance of a wheeled humanoid robot according to claim 1, characterized in that: The robot also includes a memory, a processor, and a path planning and obstacle avoidance program of the robot stored in the memory and executable on the processor, wherein the path planning and obstacle avoidance program of the robot is configured to implement the steps of the obstacle avoidance method of the industrial robot as described in any one of claims 1 to 8.

Images