Collision detection method and device for claying mechanical arm

Abstract

The invention provides a collision detection method and device for a claying mechanical arm, and relates to the technical field of coke oven slurry smearing robots, and the method comprises the following steps: S1, obtaining a movement environment data set, and calculating according to the movement environment data set to obtain a modeling parameter set of the claying mechanical arm and a wall body; s2, a modeling parameter and dynamic radius mapping table is obtained, envelope modeling is conducted on the claying mechanical arm and the wall according to the modeling parameter and dynamic radius mapping table and the modeling parameter set, and a bounding box set is obtained; the type of the bounding box comprises a sphere bounding box and a capsule bounding box; s3, selecting two bounding boxes from the bounding box set according to a preset sequence; if the two bounding boxes are both spherical bounding boxes, first-class collision detection is carried out on the two bounding boxes; if one of the two bounding boxes is a sphere bounding box and the other one is a capsule bounding box, carrying out second-class collision detection on the two bounding boxes; if the two bounding boxes are capsule bounding boxes, carrying out third-class collision detection on the two bounding boxes; obtaining a collision detection result; and S4, operation of the claying mechanical arm is controlled according to the collision detection result. By setting three different types of collision detection, high-precision real-time collision detection of the claying mechanical arm is realized on the premise of not increasing hardware complexity, and the safety and reliability of claying operation are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of coke oven mud smearing robots, and in particular to a collision detection method and device for a mud smearing robot arm. Background Art

[0002] During coke oven construction, traditional slurry application requires workers to work for extended periods within the high-temperature, cramped interior of the oven. This working environment not only poses safety risks, but also, due to the limited space, it's difficult for workers to maintain a stable working posture, leading to uneven slurry application and poorly sealed joints. In particular, the lack of manual slurry application precision in critical areas like the coke oven combustion chamber directly impacts the oven's airtightness and service life. Automating this process with a slurry application robot effectively addresses the low efficiency and inconsistent quality of manual slurry application, while also protecting workers from harsh environmental conditions.

[0003] To ensure the accuracy and safety of the plastering robot arm's operations, it was crucial from the outset to ensure that the robot arm would not collide with the furnace structure or other equipment while maneuvering within the confined space of the coke oven. Currently, the main methods for robot arm collision detection include sensor detection, image recognition detection, and envelope modeling detection.

[0004] The sensor detection method involves installing a distance sensor (such as an ultrasonic or infrared sensor) on the robotic arm to monitor the distance between the robotic arm and surrounding obstacles in real time to determine whether a collision has occurred. Alternatively, torque or current sensors can be used to indirectly detect collisions by monitoring the forces acting on the robotic arm's joints. However, this method has significant drawbacks: ultrasonic sensors significantly reduce detection accuracy in the high-temperature and dusty coke oven environment, and require high directional accuracy. Torque / current detection, on the other hand, is a passive collision detection method, triggering protection only when an actual collision occurs, resulting in abnormal torque or current, making preventive obstacle avoidance impossible. Furthermore, the sensor's poor stability in high-temperature environments increases system maintenance costs.

[0005] Image recognition detection involves deploying a vision system at the job site, using cameras to capture the robot's motion trajectory and applying image processing algorithms to identify potential collision risks. However, this approach has limited applicability in coke oven construction scenarios. Firstly, the dim lighting and high dust content inside the coke oven severely impact image quality. Secondly, the vision system requires high computing resources, making it difficult to meet real-time requirements. Furthermore, it cannot accurately identify the uneven surface of refractory bricks, making it prone to misjudgment.

[0006] To improve the accuracy and real-time performance of collision detection, existing technologies have also proposed an envelope modeling detection method. This method establishes a simplified geometric envelope (such as a sphere, cylinder, or cuboid) for each link of the robot arm and uses the separating axis theorem (SAT) to calculate the interference between the envelopes to determine collisions. However, the internal structure of the coke oven is complex, and the surfaces of the robot arm's end effector (the smearing gun head) and the furnace's refractory bricks are mostly irregular. Using a standard geometric body as the envelope will result in a large deviation between the model and the actual object. This deviation can cause two problems when working in confined spaces: first, an overly conservative envelope will significantly reduce the effective working space of the plastering robot arm; second, the simplified model may miss collision risks at certain special angles, affecting operational safety. Summary of the Invention

[0007] In order to solve the above problems, the present invention provides a collision detection method for a plastering robot arm, comprising the steps of: S1: Obtain a motion environment data set, and calculate a modeling parameter set of the plastering robot arm and the wall based on the motion environment data set; S2: Obtain a modeling parameter and dynamic radius mapping table, perform envelope modeling on the plastering robot arm and the wall according to the modeling parameter and dynamic radius mapping table and the modeling parameter set, and obtain a bounding box set; the bounding box types include spherical bounding boxes and capsule bounding boxes; S3: Select two bounding boxes from the bounding box set in a preset order; if both bounding boxes are spherical bounding boxes, perform a first-class collision test on the two bounding boxes; if one of the two bounding boxes is a spherical bounding box and the other is a capsule bounding box, perform a second-class collision test on the two bounding boxes; if both bounding boxes are capsule bounding boxes, perform a third-class collision test on the two bounding boxes; and obtain a collision test result; S4: Control the operation of the plastering robot arm according to the collision detection results.

[0008] Optionally, step S1 specifically includes: S11: Acquire a motion environment data set, the motion environment data set including: shape data and position data of each brick of the wall, and size data and angle data of each movable joint of the plastering robot arm; S12: Based on the size data and angle data of each movable joint, the intersection coordinates of the contour lines of each movable joint and the posture of each movable joint are calculated according to the kinematics forward solution algorithm; S13: fitting the contour lines of the movable joints according to the coordinates of the intersection points of the contour lines to obtain the contour lines of the movable joints; S14: fitting a contour surface according to the contour lines of each movable joint to obtain the contour surface of each movable joint; S15: The contour surface and posture of each movable joint, as well as the shape data and position data of each brick, are used as a set of modeling parameters of the plastering robot arm and the wall.

[0009] Optionally, step S2 specifically includes: S21: Obtaining shape data and position data of each brick in the modeling parameter set, obtaining the dynamic radius of each brick from the modeling parameter and dynamic radius mapping table based on the shape data and position data of each brick, constructing a geometric envelope of each brick according to the dynamic radius of each brick, performing envelope modeling on each brick using the geometric envelope of each brick, and obtaining a bounding box of each brick; S22: Obtaining the contour surface and posture of each movable joint in the modeling parameter set, obtaining the dynamic radius of each movable joint from the modeling parameter and dynamic radius mapping table based on the contour surface and posture of each movable joint, constructing the geometric envelope of each movable joint according to the dynamic radius of each movable joint, and performing envelope modeling on each movable joint using the geometric envelope of each movable joint to obtain the bounding box of each movable joint; the shapes of the geometric envelopes of each brick and each movable joint include spheres and capsules; S23: The bounding box of each brick and the bounding box of each movable joint are taken as a bounding box set.

[0010] Optionally, the first type of collision detection specifically includes: The two bounding boxes include the sphere bounding box a 1 and sphere bounding box a 2 ; Get the bounding box of the sphere a 1 The center position of the ball A 1 and dynamic radius r 1 (t) , and the sphere bounding box a 2 The center position of the ball A 2 and dynamic radius r 2 (t) , t Indicates the current time; According to the center position of the ball A 1 and the center position of the ball A 2 Calculate the first shortest distance D_1 , the calculation formula is as follows: D_1 =| A 1 ( t+ Δt )– A 2 ( t + Δt )| in, A 1 ( t + Δt ) represents the sphere bounding box a 1 At time t + Δt The center position of the ball at A 2 ( t + Δt ) represents the sphere bounding box a 2 At time t + Δt The center position of the ball at Δt is the preset time interval, |·| means obtaining the straight-line distance between the two sphere center positions; According to the dynamic radius r 1 (t) and dynamic radius r 2 (t) Calculate the safety margin ε , the calculation formula is as follows: ε= max( E ,0.1( r 1 (t) + r 2 (t) )) in, E is the preset length; If the first shortest distance D_1 <( r 1 (t) + r 2 (t) )+ ε The collision detection result is that a collision occurs, otherwise the collision detection result is that no collision occurs.

[0011] Optionally, the second type of collision detection specifically includes: The two bounding boxes include the sphere bounding box a 3 and capsule bounding box C 3 ; Get the bounding box of the sphere a3 The center position of the ball A 3 and dynamic radius r 3 (t) , and the capsule bounding box C 3 The center position of one hemisphere B 1 , the center position of the other hemisphere B 2 and the dynamic radius of the two hemispheres R 1 (t) , t Indicates the current time; Through the center of the ball A 3 , ball center position B 1 , ball center position B 2 , dynamic radius r 3 (t) and dynamic radius R 1 (t) Calculate the second shortest distance D_2 ; Constructing the capsule bounding box C 3 The Bézier curve model is used to predict the center position of the ball. B 1 and the center position of the ball B 2 After a preset time interval Δt After the position, get the predicted ball center position B 1 ' and predicted ball center position B 2 '; According to the predicted center position B 1 ' and predicted ball center position B 2 Calculate the relative linear velocity v_rel ; According to the second shortest distance D_2 , preset time interval Δt and relative linear velocity v_rel Calculate the second shortest distance after adjustment D_2' , the calculation formula is as follows: D_2' = D_2 - v_rel · Δt If the second shortest distance after adjustment D_2' <0, the collision detection result is that a collision occurs, otherwise the collision detection result is that no collision occurs.

[0012] Optionally, the third type of collision detection specifically includes: S31: Two bounding boxes including the capsule bounding box C 1 and capsule bounding box C 2 , get the capsule bounding box C 1 The center position of one hemisphere B 3 , the center position of the other hemisphere B 4 and the dynamic radius of the two hemispheres R 2 (t) , and the capsule bounding box C 2 The center position of one hemisphere B 5 , the center position of the other hemisphere B 6 and the dynamic radius of the two hemispheres R 3 (t) , t Indicates the current time; S32: Capsule bounding box C 1 and capsule bounding box C 2 Perform octree-based space division to obtain the axial bounding box U 1 and the axial bounding box U 2 ; S33: Calculate the axial bounding box U 1 With axial bounding box U 2 The minimum distance between the surfaces dist min , if the minimum distance dist min > R 2 (t) + R 3 (t) The collision detection result is that no collision occurs, otherwise the process goes to step S34; S34: Based on the hierarchical bounding box structure, the capsule bounding box C 1and capsule bounding box C 2 Each is decomposed into a cylindrical segment and two hemispheres; S35: Calculate and obtain line segments B 3 B 4 With line segment B 5 B 6 The shortest straight line distance between ; If the shortest distance is a straight line Corresponding line segments and line segments B 3 B 4 The foot of the perpendicular is on the line segment B 3 B 4 If yes, go to step S36; otherwise go to step S37; S36: Based on the shortest distance in a straight line , dynamic radius R 2 (t) and dynamic radius R 3 (t) , calculate the third shortest distance between the surfaces of the two cylindrical segments , if the third shortest distance The collision detection result is that a collision occurs, otherwise the collision detection result is that no collision occurs; S37: According to the center position of the ball B 3 , ball center position B 4 , dynamic radius R 2 (t) , ball center position B 5 , ball center position B 6 and dynamic radius R 3 (t) , calculate the shortest distance between the hemisphere and the surface of the cylindrical segment and the shortest distance between the surfaces of a hemisphere and a hemisphere ; The shortest distance between the hemisphere and the surface of the cylindrical segment and the shortest distance between the surfaces of a hemisphere and a hemisphere The minimum value among the four shortest distances is D_4 ; If the fourth shortest distance D_4≤0The collision detection result is that a collision occurs, otherwise the collision detection result is that no collision occurs.

[0013] Optionally, step S4 specifically includes: If the collision detection result indicates that a collision occurs, the plastering robot arm is controlled to stop running and the process ends; otherwise, the process returns to step S3 and reselects two bounding boxes from the bounding box set in a preset order to perform corresponding collision detection.

[0014] The present invention also provides a collision detection device for a plastering robot arm, which is used to implement the collision detection method for the plastering robot arm. The device includes: A modeling parameter acquisition module is used to obtain a motion environment data set and calculate a modeling parameter set of the plastering robot arm and the wall based on the motion environment data set; A bounding box construction module is used to obtain a modeling parameter and dynamic radius mapping table, and to perform envelope modeling of the plastering robot arm and the wall based on the modeling parameter and dynamic radius mapping table and the modeling parameter set to obtain a bounding box set; the types of bounding boxes include spherical bounding boxes and capsule bounding boxes; The collision detection module is used to select two bounding boxes from the bounding box set in a preset order; if both bounding boxes are spherical bounding boxes, perform a first-class collision detection on the two bounding boxes; if one of the two bounding boxes is a spherical bounding box and the other is a capsule bounding box, perform a second-class collision detection on the two bounding boxes; if both bounding boxes are capsule bounding boxes, perform a third-class collision detection on the two bounding boxes; and obtain a collision detection result; The plastering robot arm control module is used to control the operation of the plastering robot arm according to the collision detection results.

[0015] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the collision detection method for the plastering robot arm is implemented when the processor executes the program.

[0016] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the collision detection method of the plastering robot arm is implemented.

[0017] The present invention has the following beneficial effects: 1. Based on the shape and position data of each brick, as well as the contour surface and posture of each movable joint, the corresponding dynamic radius is obtained from the modeling parameter and dynamic radius mapping table. A more accurate geometric envelope is constructed through the dynamic radius, achieving more accurate geometric modeling and improving the matching degree between the bounding box and the bricks and movable joints, thereby improving the collision detection accuracy of the bounding box; 2. For the three situations of sphere and sphere, sphere and capsule, and capsule and capsule, the first type of collision detection, the second type of collision detection and the third type of collision detection are proposed respectively; for the first type of collision detection between sphere bounding boxes and sphere bounding boxes, the center positions of the two sphere bounding boxes are updated according to the preset time interval, and a safety margin is introduced; for the second type of collision detection between sphere bounding boxes and capsule bounding boxes, the center position of the capsule bounding box is predicted according to the Bézier curve model; for the third type of collision detection between capsule bounding boxes and capsule bounding boxes, it is divided into two situations: both perpendicular feet are located within their respective line segments of the capsule body and any perpendicular foot is located outside the line segment of the capsule body, and corresponding collision detection is performed according to different positional relationships; by setting three different types of collision detection, high-precision real-time plastering robot arm collision detection is achieved without increasing the complexity of hardware, which significantly improves the safety and reliability of plastering operations. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 The present invention is a flow chart of a collision detection method for a plastering robot arm; Figure 2 This is the structural diagram of the UR10 robotic arm; Figure 3 Schematic diagram of the bounding box of bricks and movable joints; Figure 4 This is a schematic diagram of the collision detection process of the plastering robot arm; Figure 5 This is a schematic diagram of the first type of collision detection between spherical bounding boxes; Figure 6 Schematic diagram of the second type of collision detection between the sphere bounding box and the capsule bounding box; Figure 7 Schematic diagram of the third type of collision detection between capsule bounding boxes; Figure 8 A structural diagram of an electronic device; Description of the accompanying drawings: 1-UR10 robotic arm, 2-micro ball valve, 3-mud pipe, 4-pipe winding machine, 5-single screw pump, 6-mud tank, 7-first type refractory bricks, 8-second type refractory bricks, 9-spherical bounding box of first type refractory bricks, 10-capsule bounding box of second type refractory bricks, 11-first type movable joint, 12-second type movable joint, 13-capsule bounding box of first type movable joint, 14-spherical bounding box of second type movable joint.

[0019] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings and in conjunction with the embodiments. DETAILED DESCRIPTION

[0020] The following will be combined with the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the described embodiments are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments of this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0021] Reference Figure 1 The present invention provides a collision detection method for a plastering robot arm, comprising the steps of: S1: Obtain a motion environment data set, and calculate a modeling parameter set of the plastering robot arm and the wall based on the motion environment data set; In some embodiments: Step S1 specifically includes: S11: Acquire a motion environment data set, the motion environment data set including: shape data and position data of each brick of the wall, and size data and angle data of each movable joint of the plastering robot arm; In some embodiments, the plastering robot arm of the present invention adopts the UR10 robot arm 1, and the structure of the UR10 robot arm 1 is as follows: Figure 2 As shown, its parameters are as follows: The robot uses a six-degree-of-freedom collaborative robotic arm with a working radius of 1300mm, a repeatability of ±0.1mm, and a maximum end-load of 5kg. Each joint is equipped with a dual encoder system to achieve fully closed-loop control.

[0022] The mud pipeline 3 is used to transport mud.

[0023] The pipe winding machine 4 is used to wind and place the mud pipe 3.

[0024] The single screw pump 5 is made of stainless steel, with an adjustable flow range of 0.5-10L / min and is equipped with a pressure sensor (range 0-1MPa) to achieve closed-loop control of mud flow.

[0025] The micro ball valve 2 adopts a high-temperature resistant PTFE sealing structure, has a diameter of DN10, a response time of <100ms, and communicates with the main control via the ModbusRTU protocol.

[0026] The mud tank 6 is used to store the mud used for coating.

[0027] Smear gun head assembly: A specially designed flat nozzle (outlet size 8×30mm) with an internal spiral guide groove ensures uniform mud discharge. The gun head adopts a quick-release structure for easy maintenance and replacement.

[0028] The collaborative control module, developed based on the ROS architecture, includes: (1) Motion planning submodule: uses the improved RRT algorithm for path planning; (2) Mud control submodule: realizes multi-parameter PID control of flow-pressure-valve position; (3) Safety monitoring submodule: processes collision detection signals in real time and executes safety strategies.

[0029] The working process of the UR10 robotic arm can be described as: In step 1, the UR10 robotic arm initiates a self-test, verifying the zero position of each joint using the dual encoder system. Simultaneously, the motion planning submodule loads the preset wall coordinate system parameters. The single-screw pump performs a no-load test run, the pressure sensor completes zero-point calibration, and the mud control submodule sets the initial flow rate. Finally, the micro ball valve performs an opening and closing test to ensure system readiness.

[0030] In step two, the UR10 robotic arm plans the optimal path based on the RRT algorithm and moves the coating gun head assembly precisely to the coating starting position; the collaborative control module sends a command to open the micro ball valve and simultaneously starts the single-screw pump to pre-supply slurry in low-speed mode; the mud flows through the spiral guide groove inside the gun head to fill the cavity, and the pressure sensor monitors and confirms in real time that the pipeline pressure reaches the threshold. Then the system enters standby mode and prepares to start operation.

[0031] In step three, the UR10 robotic arm moves smoothly according to the planned trajectory. The motion planning submodule adjusts the end posture in real time to keep the gun head at an inclination angle and a constant distance from the wall. The mud control submodule dynamically adjusts the speed of the single-screw pump and, combined with the pressure sensor feedback data, accurately controls the opening of the micro ball valve through a multi-parameter PID algorithm to ensure stable mud flow. The flat nozzle maintains a uniform movement, and the internal spiral guide groove allows the mud to be discharged evenly, forming an 8×30mm strip coating, achieving high-quality mud plastering operations.

[0032] In step 4, the UR10 robotic arm immediately triggers the shutdown sequence after completing the last segment of the trajectory. The micro ball valve is activated first to cut off the mud flow. The single-screw pump is delayed for 1 second to stop and release the residual pressure in the pipeline. The pressure sensor stops running after confirming that the value has returned to zero. The robotic arm automatically moves the gun head assembly to the cleaning station. The safety monitoring submodule continuously detects the environmental status. After ensuring there is no collision risk, the system enters a complete stop state.

[0033] S12: Based on the size data and angle data of each movable joint, the intersection coordinates of the contour lines of each movable joint and the posture of each movable joint are calculated according to the kinematics forward solution algorithm; In some embodiments, based on the machining dimensions of each joint arm component and after assembly in the plastering robot arm system, the actual angle measurement data of each joint of the robot arm is used as the basic data source, and the actual coordinates of the intersection points of the contour lines of each joint of the plastering robot are obtained through the robot kinematics forward solution algorithm.

[0034] By setting index requirements for the machining accuracy and actual measurement angle accuracy of each module, the actual coordinate accuracy error of the intersection point of the system joint contour lines can be controlled to be less than 1mm, thereby meeting the system collision detection modeling requirements.

[0035] S13: fitting the contour lines of the movable joints according to the coordinates of the intersection points of the contour lines to obtain the contour lines of the movable joints; S14: fitting a contour surface according to the contour lines of each movable joint to obtain the contour surface of each movable joint; S15: The contour surface and posture of each movable joint, as well as the shape data and position data of each brick, are used as a set of modeling parameters of the plastering robot arm and the wall.

[0036] S2: Obtain a modeling parameter and dynamic radius mapping table, perform envelope modeling on the plastering robot arm and the wall according to the modeling parameter and dynamic radius mapping table and the modeling parameter set, and obtain a bounding box set; the bounding box types include spherical bounding boxes and capsule bounding boxes; In some embodiments, the prior art typically uses volume modeling (regular shapes such as cylinders, spheres, and rectangles) to establish bounding boxes. The volume modeling detection method selects one or more discrete points in a joint, creates a cylindrical, spherical, or rectangular volume, and then determines whether a collision has occurred by calculating the distance between the center of the sphere or volume and the rest of the joint. However, most actual joints are irregular shapes, and using a regular volume model to calculate the distance does not fit the actual shape of the robotic arm. Consequently, collision detection using a volume often results in misjudgments.

[0037] For example, when a cylindrical bounding box is used to envelop a robotic arm joint, there is often an obvious envelope mismatch problem at the joint connection. Since robotic arm joints usually have irregular geometric shapes and complex motion characteristics, a simple cylindrical model is difficult to accurately fit the actual structure, resulting in poor collision detection results. In terms of geometry, joints often have flanges, grooves or asymmetric designs, and a cylindrical bounding box can only be enveloping with a fixed diameter and axial range, resulting in a large gap between the joint edge and the enveloping volume. These uncovered areas are easily overlooked during collision detection, resulting in the system being unable to accurately identify potential interference risks.

[0038] Furthermore, robotic arm joints often involve multiple degrees of freedom (DOFs) such as rotation and deflection during motion, and static cylindrical envelopes cannot dynamically adjust to the joint's posture. As the joint rotates, the deviation between the envelope and the actual contour increases, potentially leading to two types of false detections: false interference between the envelope and surrounding structures, resulting in a false collision; and actual contact obscured by the rigid boundaries of the envelope, leading to missed detections.

[0039] The limitations of this enveloping method are particularly prominent in high-speed or precision operation scenarios. It not only reduces the reliability of collision detection, but may also affect the overall motion performance of the robotic arm.

[0040] In the existing technology, in order to solve the error problem of envelope modeling, some solutions adopt the method of adding sensors or visual detection devices to the robotic arm. However, actual applications have shown that this type of method still has inherent defects: the visual detection system is limited by the camera's viewing angle, and its detection effect will fluctuate significantly with the change of the angle between the robotic arm's movement direction and the lens optical axis. In particular, when the robotic arm and the camera are in a vertical position, the system will completely lose the ability to perceive collisions. Similarly, contact detection solutions such as position sensors and force sensors can indirectly infer the risk of collision through position information or torque feedback, but due to factors such as signal transmission delay and mechanical vibration interference, it is still difficult to achieve accurate and real-time collision avoidance. These sensor-based detection methods cannot fundamentally eliminate the hidden dangers of robotic arm collisions.

[0041] At their core, these technical bottlenecks stem from the inherent flaws of sensor-dependent detection methods. Current sensor-based collision detection systems face dual constraints: First, the sensors must be integrated within the robotic arm, and their measurement benchmarks are inevitably affected by the cumulative errors of mechanical assembly. Second, the sensors' inherent measurement accuracy limitations and lack of environmental adaptability further constrain the accuracy of the detection model. This systematic error, introduced at the source of data acquisition, makes it difficult for modeling methods relying on sensor data to break through the accuracy ceiling.

[0042] The technical solution proposed in this paper overcomes the limitations of bounding boxes based on conventional regular geometric shapes (cylinders, spheres, and cuboids). It innovatively employs a sphere-capsule hybrid bounding strategy and introduces a dynamic radius adjustment mechanism. This allows for dynamic radius optimization by establishing a mapping table between modeling parameters and dynamic radius. This solution, through a multi-level geometric approximation algorithm and an octree-based spatial pre-segmentation strategy, improves the matching accuracy between the collision detection model and the actual contour of the robotic arm by over 40%. The specific process is as follows: Step S2 specifically includes: S21: Obtaining shape data and position data of each brick in the modeling parameter set, obtaining the dynamic radius of each brick from the modeling parameter and dynamic radius mapping table based on the shape data and position data of each brick, constructing a geometric envelope of each brick according to the dynamic radius of each brick, performing envelope modeling on each brick using the geometric envelope of each brick, and obtaining a bounding box of each brick; S22: Obtaining the contour surface and posture of each movable joint in the modeling parameter set, obtaining the dynamic radius of each movable joint from the modeling parameter and dynamic radius mapping table based on the contour surface and posture of each movable joint, constructing the geometric envelope of each movable joint according to the dynamic radius of each movable joint, and performing envelope modeling on each movable joint using the geometric envelope of each movable joint to obtain the bounding box of each movable joint; the shapes of the geometric envelopes of each brick and each movable joint include spheres and capsules; In some embodiments, the bounding boxes of bricks and movable joints are as follows: Figure 3 As shown, for the first type of refractory brick 7 having a shape close to a sphere, a spherical bounding box 9 of the first type of refractory brick is constructed; for the second type of refractory brick 8 having a shape close to a capsule, a capsule bounding box 10 of the second type of refractory brick is constructed; for the first type of movable joint 11 having a shape close to a capsule, a capsule bounding box 13 of the first type of movable joint is constructed; for the second type of movable joint 12 having a shape close to a sphere, a spherical bounding box 14 of the second type of movable joint is constructed; S23: The bounding box of each brick and the bounding box of each movable joint are taken as a bounding box set.

[0043] S3: Select two bounding boxes from the bounding box set in a preset order; if both bounding boxes are spherical bounding boxes, perform a first-class collision test on the two bounding boxes; if one of the two bounding boxes is a spherical bounding box and the other is a capsule bounding box, perform a second-class collision test on the two bounding boxes; if both bounding boxes are capsule bounding boxes, perform a third-class collision test on the two bounding boxes; and obtain a collision test result; In some embodiments, the collision detection process of the plastering robot arm is as follows: Figure 4 As shown in the figure, the collision detection process of the mud-smearing robot arm mainly faces three types of collision detection problems, namely, between spheres and spheres, between spheres and capsules, and between capsules and capsules.

[0044] In some embodiments, the first type of collision detection between sphere bounding boxes is as follows: Figure 5 As shown; The first type of collision detection specifically includes: The two bounding boxes include the sphere bounding box a 1 and sphere bounding box a 2 ; Get the bounding box of the sphere a 1 The center position of the ball A 1 and dynamic radius r 1 (t) , and the sphere bounding box a 2 The center position of the ball A 2 and dynamic radius r 2 (t) , t Indicates the current time; In some embodiments, the dynamic radius parameter is dynamically adjusted through a pre-built joint motion trajectory database, and a mapping table of modeling parameters and dynamic radius is established. r 1 (t) The expression is as follows: r 1 (t) = r 1-base + k·|θ'(t)| in θ'(t) is the joint angular velocity, k is the dynamic compensation coefficient, r 1-base It is the corresponding length data obtained from the modeling parameter and dynamic radius mapping table.

[0045] According to the center position of the ball A 1 and the center position of the ball A 2 Calculate the first shortest distance D_1 , the calculation formula is as follows: D_1 =| A 1 ( t + Δt )– A 2 ( t + Δt )| in, A 1 ( t + Δt ) represents the sphere bounding box a 1 At time t + Δt The center position of the ball at A 2 ( t+ Δt ) represents the sphere bounding box a 2 At time t + Δt The center position of the ball at Δt is the preset time interval, |·| means obtaining the straight-line distance between the two sphere center positions; In some embodiments, the introduction Δt Center position of the ball A 1 and the center position of the ball A 2 Make speed predictions, Δt =0.1 T ( T is the control cycle); According to the dynamic radius r 1 (t) and dynamic radius r 2 (t) Calculate the safety margin ε , the calculation formula is as follows: ε= max( E ,0.1( r 1 (t) + r 2 (t) )) in, E is the preset length; In some embodiments, E Take 5mm; If the first shortest distance D_1 <( r 1 (t) + r 2 (t) )+ ε The collision detection result is that a collision occurs, otherwise the collision detection result is that no collision occurs.

[0046] In some embodiments, the second type of collision detection between the sphere bounding box and the capsule bounding box is as follows: Figure 6 As shown; The second type of collision detection specifically includes: The two bounding boxes include the sphere bounding box a 3 and capsule bounding box C 3 ; Get the bounding box of the sphere a 3The center position of the ball A 3 and dynamic radius r 3 (t) , and the capsule bounding box C 3 The center position of one hemisphere B 1 , the center position of the other hemisphere B 2 and the dynamic radius of the two hemispheres R 1 (t) , t Indicates the current time; Through the center of the ball A 3 , ball center position B 1 , ball center position B 2 , dynamic radius r 3 (t) and dynamic radius R 1 (t) Calculate the second shortest distance D_2 ; In some embodiments, the second shortest distance between the two D_2 There are two situations: A 3 、 B 1 and B 2 The three points are not collinear, such as Figure 6 (a) As shown; the three points are collinear, such as Figure 6 (b) shown.

[0047] 1. The calculation steps for the first case of sphere and capsule (three points are not collinear) are as follows: Establish a passing point A 3 And with the straight line B 1 B 2 Orthogonal planes, record the intersection points P 1 For the orthogonal projection point, through the parameterized equation =( B 1 P 1 ·B 1 B 2 ) / | B1 B 2 |²Confirm P 1 exist B 1 B 2 Relative position on: like ∈[0,1], then P 1 Located within the cylindrical segment; like <0, P 1 near B 1 endpoints; like >1, P 1 near B 2 endpoints; when P 1 When inside the capsule segment: second shortest distance D_2 is the distance from the center of the sphere to the axis, which can be expressed as max(0,| A 3 P 1 |-( r 3 (t) + R 1 (t) )); when P 1 When outside the capsule endpoint: take the distance P 1 The nearest endpoint is calculated as B 1 For example, the second shortest distance D_2 =max(0,| A 3 B 1 |-( r 3 (t) + R 1 (t) )); 2. The calculation steps for the second case of the sphere and capsule (three points collinear) are as follows: Let the center of the sphere be A 3 To capsule point B 1The distance is D_2.1 ,point Arrive B 2 The distance is D_2.2 . Take the minimum value according to the following formula: D_2 =min( D_2.1,D_2.2 ) Constructing the capsule bounding box C 3 The Bézier curve model is used to predict the center position of the ball. B 1 and the center position of the ball B 2 After a preset time interval Δt After the position, get the predicted ball center position B 1 ' and predicted ball center position B 2 '; In some embodiments, motion continuity optimization is performed using a Bézier curve model to predict the next cycle endpoint position: B 1 '= B 1 +3( T2 - T1 ) / Δt 、 B 2 '= B 2 +3( T2 - T1 ) / Δt in, T1 , T2 is the position of the first two cycles; According to the predicted center position B 1 ' and predicted ball center position B 2 Calculate the relative linear velocity v_rel ; According to the second shortest distance D_2 , preset time interval Δt and relative linear velocity v_rel Calculate the second shortest distance after adjustment D_2' , the calculation formula is as follows: D_2' = D_2 - v_rel · Δt If the second shortest distance after adjustment D_2'<0, the collision detection result is that a collision occurs, otherwise the collision detection result is that no collision occurs.

[0048] In some embodiments, the third type of collision detection between capsule bounding boxes is as follows: Figure 7 As shown; The third type of collision detection specifically includes: S31: Two bounding boxes including the capsule bounding box C 1 and capsule bounding box C 2 , get the capsule bounding box C 1 The center position of one hemisphere B 3 , the center position of the other hemisphere B 4 and the dynamic radius of the two hemispheres R 2 (t) , and the capsule bounding box C 2 The center position of one hemisphere B 5 , the center position of the other hemisphere B 6 and the dynamic radius of the two hemispheres R 3 (t) , t Indicates the current time; S32: Capsule bounding box C 1 and capsule bounding box C 2 Perform octree-based space division to obtain the axial bounding box U 1 and the axial bounding box U 2 ; S33: Calculate the axial bounding box U 1 With axial bounding box U 2 The minimum distance between the surfaces dist min , if the minimum distance dist min > R 2 (t) + R 3 (t) The collision detection result is that no collision occurs, otherwise the process goes to step S34; S35: Calculate and obtain line segments B3 B 4 With line segment B 5 B 6 The shortest straight line distance between ; If the shortest distance is a straight line Corresponding line segments and line segments B 3 B 4 The foot of the perpendicular is on the line segment B 3 B 4 Within, and the shortest distance in a straight line Corresponding line segments and line segments B 5 B 6 The foot of the perpendicular is on the line segment B 5 B 6 If yes, go to step S36; otherwise go to step S37; In some embodiments, calculating two line segments B 3 B 4 and B 5 B 6 The shortest straight line distance:

[0049] in For line segments B 3 B 4 , line segment B 5 B 6 The direction vector, is the line vector connecting any two endpoints. If the two line segments are parallel ( ), calculated using the perpendicular projection method.

[0050] Line segment B 3 B 4 With line segment B 5 B 6 Shortest distance to a straight line The foot of the corresponding line segments are all located within their respective line segments. Figure 7 As shown in (a), the situation where any perpendicular foot is outside the line segment is as follows Figure 7 As shown in (b), the position of the vertical foot is P2; S36: Based on the shortest distance in a straight line , dynamic radius R 2 (t) and dynamic radius R 3 (t) , calculate the third shortest distance between the surfaces of the two cylindrical segments , if the third shortest distance The collision detection result is that a collision occurs, otherwise the collision detection result is that no collision occurs; In some embodiments, the third shortest distance The expression is:

[0051] S37: According to the center position of the ball B 3 , ball center position B 4 , dynamic radius R 2 (t) , ball center position B 5 , ball center position B 6 and dynamic radius R 3 (t) , calculate the shortest distance between the hemisphere and the surface of the cylindrical segment and the shortest distance between the surfaces of a hemisphere and a hemisphere ; The shortest distance between the hemisphere and the surface of the cylindrical segment and the shortest distance between the surfaces of a hemisphere and a hemisphere The minimum value among the four shortest distances is D_4 ; If the fourth shortest distance D_4≤0 The collision detection result is that a collision occurs, otherwise the collision detection result is that no collision occurs.

[0052] In some embodiments, for a capsule whose foot is beyond the line segment, the nearest sphere center position (e.g. C 2 The center position of the ball B 6 ); Calculate the center position of the sphere B 6 To another capsule C 1 The closest distance to the axis: (1) If the center of the ball isB 6 exist C 1 In cylindrical segment projection: ,in Q proj The center position of the ball B 6 In the capsule C 1 Projection of the axis.

[0053] (2) If the center of the ball is B 6 near C 1 The endpoint hemisphere of: ,in Q nearest for C 1 Position from the center of the ball B 6 The nearest center of the ball, here is the center of the ball B 4 ; D_4 =min( D proj , D end ).

[0054] S4: Control the operation of the plastering robot arm according to the collision detection results.

[0055] In some embodiments: Step S4 specifically includes: If the collision detection result indicates that a collision occurs, the plastering robot arm is controlled to stop running and the process ends; otherwise, the process returns to step S3 and reselects two bounding boxes from the bounding box set in a preset order to perform corresponding collision detection.

[0056] In some embodiments, when the safety monitoring submodule detects a collision signal or system abnormality, it immediately freezes all movements of the UR10 robotic arm and forcibly closes the micro ball valve to block the mud flow; the single-screw pump automatically switches to safety mode, maintaining mud flow at a continuous low speed to prevent pipeline blockage while waiting for manual intervention. The operator can choose to resume the operation process or perform an emergency shutdown according to the actual situation to ensure the safety of equipment and personnel.

[0057] The present invention also provides a collision detection device for a plastering robot arm, which is used to implement the collision detection method for the plastering robot arm. The device includes: A modeling parameter acquisition module is used to obtain a motion environment data set and calculate a modeling parameter set of the plastering robot arm and the wall based on the motion environment data set; A bounding box construction module is used to obtain a modeling parameter and dynamic radius mapping table, and to perform envelope modeling of the plastering robot arm and the wall based on the modeling parameter and dynamic radius mapping table and the modeling parameter set to obtain a bounding box set; the types of bounding boxes include spherical bounding boxes and capsule bounding boxes; The collision detection module is used to select two bounding boxes from the bounding box set in a preset order; if both bounding boxes are spherical bounding boxes, perform a first-class collision detection on the two bounding boxes; if one of the two bounding boxes is a spherical bounding box and the other is a capsule bounding box, perform a second-class collision detection on the two bounding boxes; if both bounding boxes are capsule bounding boxes, perform a third-class collision detection on the two bounding boxes; and obtain a collision detection result; The plastering robot arm control module is used to control the operation of the plastering robot arm according to the collision detection results.

[0058] In some embodiments, see Figure 8 , Figure 8 The electronic device 80 provided in the embodiment of the present application includes a memory 81 and a processor 82; the memory 81 stores a computer program, wherein the computer program, when executed by the processor, implements the collision detection method for the plastering robot arm.

[0059] Specifically, the processor 82 may include, for example, a general-purpose microprocessor, an instruction set processor and / or a related chipset and / or a dedicated microprocessor (e.g., an application-specific integrated circuit (ASIC)), etc. The processor 82 may also include onboard memory for caching purposes. The processor 82 may be a single processing unit or multiple processing units for executing different actions of the method flow according to the embodiments of the present application.

[0060] Memory 81 can be, for example, any medium capable of containing, storing, conveying, disseminating, or transmitting instructions. For example, memory 81 can include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, components, or communication media. Specific examples of memory 81 include: magnetic storage devices, such as magnetic tape or hard disk drives (HDDs); optical storage devices, such as compact discs (CD-ROMs); random access memory (RAM) or flash memory; and / or wired or wireless communication links.

[0061] This application also provides a computer-readable medium storing a computer program that, when executed by a processor, implements the aforementioned collision detection method for a plastering robot arm. This computer-readable medium may be included in the device / apparatus / system described in the aforementioned embodiments, or it may exist independently and not incorporated into the device / apparatus / system. The aforementioned computer-readable medium carries one or more programs that, when executed, implement the methods described in the embodiments of this application.

[0062] According to an embodiment of the present application, a computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium or any combination thereof. A computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or component, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present application, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, device, or component. In the present application, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, which carries computer-readable program code. Such a propagated data signal may take a variety of forms, including but not limited to an electromagnetic signal, an optical signal, or any suitable combination thereof. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium that can transmit, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code embodied on a computer-readable medium may be transmitted using any suitable medium, including but not limited to wireless, wired, optical cable, radio frequency signals, or any suitable combination thereof.

[0063] Those skilled in the art will understand that the features described in the various embodiments and / or claims of the present application may be combined and / or combined in a variety of ways, even if such combinations or combinations are not explicitly described in the present application. In particular, without departing from the spirit and teachings of the present application, the features described in the various embodiments and / or claims of the present application may be combined and / or combined in a variety of ways. All of these combinations and / or combinations fall within the scope of the present application. Therefore, the scope of the present application should not be limited to the above-mentioned embodiments, but should be determined not only by the attached claims, but also by the equivalents of the attached claims. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application should be included in the scope of protection of the present application.

Claims

1. A collision detection method for a plastering robot arm, characterized in that: Including steps: S1: Obtain a motion environment data set, and calculate a modeling parameter set of the plastering robot arm and the wall based on the motion environment data set; S2: Obtain a modeling parameter and dynamic radius mapping table, perform envelope modeling on the plastering robot arm and the wall according to the modeling parameter and dynamic radius mapping table and the modeling parameter set, and obtain a bounding box set; the bounding box types include spherical bounding boxes and capsule bounding boxes; S3: Select two bounding boxes from the bounding box set in a preset order; if both bounding boxes are spherical bounding boxes, perform a first-class collision test on the two bounding boxes; if one of the two bounding boxes is a spherical bounding box and the other is a capsule bounding box, perform a second-class collision test on the two bounding boxes; if both bounding boxes are capsule bounding boxes, perform a third-class collision test on the two bounding boxes; and obtain a collision test result; S4: Control the operation of the plastering robot arm according to the collision detection results.

2. The collision detection method for a plastering robot arm according to claim 1, characterized in that: Step S1 specifically includes: S11: Acquire a motion environment data set, the motion environment data set including: shape data and position data of each brick of the wall, and size data and angle data of each movable joint of the plastering robot arm; S12: Based on the size data and angle data of each movable joint, the intersection coordinates of the contour lines of each movable joint and the posture of each movable joint are calculated according to the kinematics forward solution algorithm; S13: fitting the contour lines of the movable joints according to the coordinates of the intersection points of the contour lines to obtain the contour lines of the movable joints; S14: fitting a contour surface according to the contour lines of each movable joint to obtain the contour surface of each movable joint; S15: The contour surface and posture of each movable joint, as well as the shape data and position data of each brick, are used as a set of modeling parameters of the plastering robot arm and the wall.

3. The collision detection method for a plastering robot arm according to claim 1, characterized in that: Step S2 specifically includes: S21: Obtaining shape data and position data of each brick in the modeling parameter set, obtaining the dynamic radius of each brick from the modeling parameter and dynamic radius mapping table based on the shape data and position data of each brick, constructing a geometric envelope of each brick according to the dynamic radius of each brick, performing envelope modeling on each brick using the geometric envelope of each brick, and obtaining a bounding box of each brick; S22: Obtaining the contour surface and posture of each movable joint in the modeling parameter set, obtaining the dynamic radius of each movable joint from the modeling parameter and dynamic radius mapping table based on the contour surface and posture of each movable joint, constructing the geometric envelope of each movable joint according to the dynamic radius of each movable joint, and performing envelope modeling on each movable joint using the geometric envelope of each movable joint to obtain the bounding box of each movable joint; the shapes of the geometric envelopes of each brick and each movable joint include spheres and capsules; S23: The bounding box of each brick and the bounding box of each movable joint are taken as a bounding box set.

4. The collision detection method for a plastering robot arm according to claim 1, characterized in that: The first type of collision detection specifically includes: The two bounding boxes include the sphere bounding box a 1 and sphere bounding box a 2 ; Get the bounding box of the sphere a 1 The center position of the ball A 1 and dynamic radius r 1 (t) , and the sphere bounding box a 2 The center position of the ball A 2 and dynamic radius r 2 (t) , t Indicates the current time; According to the center position of the ball A 1 and the center position of the ball A 2 Calculate the first shortest distance D_1 , the calculation formula is as follows: D_1 =| A 1 ( t + Δt )– A 2 ( t + Δt )| in, A 1 ( t + Δt ) represents the sphere bounding box a 1 At time t + Δt The center position of the ball at A 2 ( t + Δt ) represents the sphere bounding box a 2 At time t + Δt The center position of the ball at Δt is the preset time interval, |·| means obtaining the straight-line distance between the two sphere center positions; According to the dynamic radius r 1 (t) and dynamic radius r 2 (t) Calculate the safety margin ε , the calculation formula is as follows: ε= max( E ,0.1( r 1 (t) + r 2 (t) )) in, E is the preset length; If the first shortest distance D_1 <( r 1 (t) + r 2 (t) )+ ε The collision detection result is that a collision occurs, otherwise the collision detection result is that no collision occurs.

5. The collision detection method for a plastering robot arm according to claim 1, characterized in that: The second type of collision detection specifically includes: The two bounding boxes include the sphere bounding box a 3 and capsule bounding box C 3 ; Get the bounding box of the sphere a 3 The center position of the ball A 3 and dynamic radius r 3 (t) , and the capsule bounding box C 3 The center position of one hemisphere B 1 , the center position of the other hemisphere B 2 and the dynamic radius of the two hemispheres R 1 (t) , t Indicates the current time; Through the center of the ball A 3 , ball center position B 1 , ball center position B 2 , dynamic radius r 3 (t) and dynamic radius R 1 (t) Calculate the second shortest distance D_2 ; Constructing the capsule bounding box C 3 The Bézier curve model is used to predict the center position of the ball. B 1 and the center position of the ball B 2 After a preset time interval Δt After the position, get the predicted ball center position B 1 ' and predicted ball center position B 2 '; According to the predicted center position B 1 ' and predicted ball center position B 2 Calculate the relative linear velocity v_rel ; According to the second shortest distance D_2 , preset time interval Δt and relative linear velocity v_rel Calculate the second shortest distance after adjustment D_2' , the calculation formula is as follows: D_2' = D_2 - v_rel · Δt If the second shortest distance after adjustment D_2' <0, the collision detection result is that a collision occurs, otherwise the collision detection result is that no collision occurs.

6. The collision detection method for a plastering robot arm according to claim 1, characterized in that: The third type of collision detection specifically includes: S31: Two bounding boxes including the capsule bounding box C 1 and capsule bounding box C 2 , get the capsule bounding box C 1 The center position of one hemisphere B 3 , the center position of the other hemisphere B 4 and the dynamic radius of the two hemispheres R 2 (t) , and the capsule bounding box C 2 The center position of one hemisphere B 5 , the center position of the other hemisphere B 6 and the dynamic radius of the two hemispheres R 3 (t) , t Indicates the current time; S32: Capsule bounding box C 1 and capsule bounding box C 2 Perform octree-based space division to obtain the axial bounding box U 1 and the axial bounding box U 2 ; S33: Calculate the axial bounding box U 1 With axial bounding box U 2 The minimum distance between the surfaces dist min , if the minimum distance dist min > R 2 (t) + R 3 (t) The collision detection result is that no collision occurs, otherwise the process goes to step S34; S34: Based on the hierarchical bounding box structure, the capsule bounding box C 1 and capsule bounding box C 2 Each is decomposed into a cylindrical segment and two hemispheres; S35: Calculate and obtain line segments B 3 B 4 With line segment B 5 B 6 The shortest straight line distance between ; If the shortest distance is a straight line Corresponding line segments and line segments B 3 B 4 The foot of the perpendicular line segment B 3 B 4 Within, and the shortest distance in a straight line Corresponding line segments and line segments B 5 B 6 The foot of the perpendicular line segment B 5 B 6 If yes, go to step S36; otherwise go to step S37; S36: Based on the shortest distance in a straight line , dynamic radius R 2 (t) and dynamic radius R 3 (t) , calculate the third shortest distance between the surfaces of the two cylindrical segments , if the third shortest distance The collision detection result is that a collision occurs, otherwise the collision detection result is that no collision occurs; S37: According to the center position of the ball B 3 , ball center position B 4 , dynamic radius R 2 (t) , ball center position B 5 , ball center position B 6 and dynamic radius R 3 (t) , calculate the shortest distance between the hemisphere and the surface of the cylindrical segment and the shortest distance between the surfaces of a hemisphere and a hemisphere ; The shortest distance between the hemisphere and the surface of the cylindrical segment and the shortest distance between the surfaces of a hemisphere and a hemisphere The minimum value among the four shortest distances is D_4 ; If the fourth shortest distance D_4≤0 The collision detection result is that a collision occurs, otherwise the collision detection result is that no collision occurs.

7. The collision detection method for a plastering robot arm according to claim 1, characterized in that: Step S4 specifically includes: If the collision detection result indicates that a collision occurs, the plastering robot arm is controlled to stop running and the process ends; otherwise, the process returns to step S3 and reselects two bounding boxes from the bounding box set in a preset order to perform corresponding collision detection.

8. A collision detection device for a plastering robot arm, used to implement the collision detection method for a plastering robot arm according to any one of claims 1 to 7, characterized in that: The device comprises: A modeling parameter acquisition module is used to obtain a motion environment data set and calculate a modeling parameter set of the plastering robot arm and the wall based on the motion environment data set; A bounding box construction module is used to obtain a modeling parameter and dynamic radius mapping table, and to perform envelope modeling of the plastering robot arm and the wall based on the modeling parameter and dynamic radius mapping table and the modeling parameter set to obtain a bounding box set; the types of bounding boxes include spherical bounding boxes and capsule bounding boxes; The collision detection module is used to select two bounding boxes from the bounding box set in a preset order; if both bounding boxes are spherical bounding boxes, perform a first-class collision detection on the two bounding boxes; if one of the two bounding boxes is a spherical bounding box and the other is a capsule bounding box, perform a second-class collision detection on the two bounding boxes; if both bounding boxes are capsule bounding boxes, perform a third-class collision detection on the two bounding boxes; and obtain a collision detection result; The plastering robot arm control module is used to control the operation of the plastering robot arm according to the collision detection results.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the collision detection method for the plastering robotic arm according to any one of claims 1 to 7 is implemented.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the collision detection method for the plastering robot arm according to any one of claims 1 to 7 is implemented.

Images