Electrified robot motion safety planning system based on multi-performance quantitative constraint

Abstract

The invention relates to the field of live-line robot motion safety planning, and provides a live-line robot motion safety planning system based on multi-performance quantitative constraint, an object detection module analyzes and determines coordinates of multiple objects in a space by collecting RGB image information of a visual sensor and three-dimensional point cloud data collected by a laser radar sensor, and determines the coordinates of the multiple objects in the space; the global planning module receives object information and plans a global motion path of the mechanical arm through a sampling-calculation method, the local reconstruction module is in data connection with the object detection module and the global planning module and optimizes a local path, and the safety verification module carries out safety re-verification on robot motion and carries out safety re-verification on the robot motion by receiving the motion path of the mechanical arm and the object information. According to the method, the motion path of the mechanical arm is globally planned for the first time, the object change is detected in real time, and the local path is locally planned for multiple times, so that the motion planning capability of the robot on the dynamic target is effectively improved.

Description

Technical Field

[0001] The present invention relates to the field of live robot motion safety planning, in particular to a live robot motion safety planning system based on multiple performance quantification constraints. Background Art

[0002] Live working is a common maintenance method in the field of power maintenance. Live working can effectively maintain the power supply network, reduce the number of power outages, and improve power supply reliability. At the same time, since live operations are extremely dangerous, live robots are gradually being used to replace manual operations. Due to the complex live working environment and the diverse operating objects, live robot motion safety planning has become the core technology for safe operations.

[0003] Motion safety planning usually includes object recognition, motion strategy selection, and motion path planning. Currently, some research designs motion strategies by detecting the position and size information of environmental objects and work objects, analyzing the starting and ending positions of the end of the robot arm, and specifying non-collision path planning in Cartesian space to provide a feasible motion path for the end of the robot arm. However, in actual situations, environmental objects and work objects are likely to be dynamically changing, and it is difficult to handle dynamically changing objects and work objects using static planning strategies. Summary of the Invention

[0004] In order to solve the above technical problems, the present invention provides a live robot motion safety planning system based on multi-performance quantification constraints. By globally planning the motion path of the robotic arm for the first time, detecting object changes in real time, and locally planning local paths multiple times, the robot's motion path for dynamic targets is effectively planned.

[0005] A live robot motion safety planning system based on multiple performance quantification constraints includes: an object detection module for collecting RGB image information and three-dimensional point cloud data, and determining a surface coordinate set A1 of a target object in front of the robot and a surface coordinate set A2 of an obstacle object based on the RGB image information and the three-dimensional point cloud data;

[0006] The global planning module is used to receive the surface coordinate set A1 of the target object before the robot's manipulator moves, the surface coordinate set A2 of the obstacle object, and the initial coordinates of the end of the manipulator. Destination coordinates of the end of the robotic arm and safety distance d safe , and then plan the global motion path of the end of the robot's manipulator through the global path generation method;

[0007] The local reconstruction module is used to receive the surface coordinate set A1 of the target object, the surface coordinate set A2 of the obstacle object, the starting point near the target object or the obstacle object during the operation of the robot's manipulator and end point Update the local path near the target object or obstacle in the motion path that has not been passed according to the local path generation method;

[0008] The safety verification module is used to continuously receive the surface coordinate set A1 of the target object, the surface coordinate set A2 of the obstacle object, the current coordinate P of the end of the robot arm and the local path during the operation of the robot arm, so as to perform safety verification on all local paths until the robot arm moves to the target coordinate

[0009] Preferably, the object detection module extracts shape features based on RGB image information, performs contour analysis based on the shape features to determine the target object and obstacle objects, uses three-dimensional point cloud data to construct the geometric dimensions and spatial positioning of the target object and obstacle objects, and obtains the surface coordinate set A1 of the target object and the surface coordinate set A2 of the obstacle objects based on the results of the contour analysis.

[0010] Preferably, the global path generation method includes: generating a first valid sampling point set S2, obtaining the initial coordinates of the end of the mechanical arm Destination coordinates and the safety distance is d safe ,by and Construct a rectangular parallelepiped sampling space for the diagonal endpoints, uniformly distribute the sampling points with a fixed density in the sampling space to generate a sampling point set S1, and screen out the safe distance d between the target object surface and the obstacle surface safe Several valid sampling points other than , and generate a first valid sampling point set S2;

[0011] Generate global motion path, starting point Get the temporary starting point p and connect the K adjacent valid sampling points in the first valid sampling point set S2 to generate K global neighborhood graphs. Perform collision detection on the edges connected by the K global neighborhood graphs and only retain the k global neighborhood graphs with no collision edges. Determine whether the k global neighborhood graphs contain the end point. Yes, the fit includes the endpoint The global neighborhood graph is used to generate the global motion path. If not, jump to the step of determining the adjacent valid sampling points.

[0012] Determine the global effective sampling points and calculate the function values ​​F of k global neighborhood graphs respectively through the first function (O) , the function value F (O) The last valid sampling point of the minimum global neighborhood graph is defined as the temporary starting point p, and the process jumps to the step of generating the global neighborhood graph.

[0013] Preferably, the first function is:

[0014]

[0015] in, Indicates the initial coordinates of the end of the robotic arm The Euclidean distance to all valid sampling points in the connected global neighborhood graph, Represents the coordinates of all valid sampling points to target objects in the connected global neighborhood graph The Euclidean distance of .

[0016] Preferably, the local path generation method includes: generating a second valid sampling point set S3, receiving a surface coordinate set A1 of a target object and a surface coordinate set A2 of an obstacle object, defining the first valid sampling point of the global motion path near the target object or the obstacle object as the starting point The last valid sampling point is defined as the end point Based on the surface coordinate set A1 of the target object, the surface coordinate set A2 of the obstacle object, the first sampling point set S1 and the safety distance d safe Screen out the safe distance d between the target object surface and the obstacle surface safe Several valid sampling points other than , and generate a second valid sampling point set S3;

[0017] Generate local path, starting point Or the temporary starting point t is connected to a neighboring valid sampling point in the second valid sampling point set S3 to generate a local area graph, and determines whether the neighboring valid sampling point is the end point Yes, the fit includes the endpoint The local neighborhood graph is used to generate a local path. If not, jump to the step of determining the adjacent valid sampling points.

[0018] Determine the local effective sampling points and calculate the starting points respectively through the second function F between adjacent valid sampling points, temporary starting point t and adjacent valid sampling points (k) Difference, judge the two F (k) Is the difference less than the convergence threshold ε? If yes, connect the adjacent valid sampling points to the starting point Or connect the adjacent valid sampling points with the temporary starting point t, and define the adjacent valid sampling points as the new starting point t, and jump to the step of generating a local path; otherwise, the adjacent valid sampling points are not connected to the starting point Or a temporary starting point t connection to jump to the step of generating a local path.

[0019] Preferably, the second function is:

[0020]

[0021] in, Indicates the starting point The Euclidean distance to the adjacent valid sampling point or temporary starting point t, Indicates the adjacent valid sampling point or temporary starting point t to the end point The Euclidean distance of .

[0022] Preferably, the local reconstruction module includes a multi-constraint rule optimization unit for correcting the local path, including an evaluation function:

[0023] J i =(max(0,d safe -d i )*max(0,θ i -θ min ))

[0024] Among them, θ min is the minimum movement angle of the robot arm, θ i is the motion angle of the manipulator between the first i-th valid sampling point i and the i-1th valid sampling point, d i Represents the distance between the i-th valid sampling point and the target object, and J i The i-th valid sampling point with =0 is eliminated, and the remaining valid sampling points are refitted to optimize the local path.

[0025] Preferably, the safety verification module includes: determining the valid sampling points that have not passed the global motion path according to the current coordinate P of the end of the manipulator, judging whether the number of valid sampling points that have not passed is 0, if yes, the manipulator ends the movement, if not, judging whether there is a valid sampling point that has not passed the distance from the obstacle or the surface of the target object that is less than the safety distance d safe If yes, the local reconstruction module re-updates the corresponding local path. If no, it continues to receive the next set of surface coordinate sets A1 of the target object, the surface coordinate set A2 of the obstacle object, the current coordinates P of the end of the robotic arm, and the local path.

[0026] Compared with the prior art, the present invention has the following beneficial effects:

[0027] 1. The present invention constructs a Cartesian space by analyzing visual data and lidar data, maps obstacles and target objects into the Cartesian space, and realizes the accurate quantification of the surface position information of obstacles and target objects.

[0028] 2. The present invention plans the global motion path of the robot arm in space by using a global path generation method in Cartesian space, accurately locates obstacles and target objects in motion, and optimizes the paths near obstacles and target objects through a local path generation method, thereby effectively planning the robot's motion path for dynamic targets. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1It is a workflow diagram of the present invention.

[0030] Figure 2 It is a workflow diagram of the local reconstruction module in the present invention. DETAILED DESCRIPTION

[0031] The following embodiments of the present invention are described in further detail with reference to the accompanying drawings and examples. The following examples are used to illustrate the present invention but are not intended to limit the scope of the present invention.

[0032] The present invention provides a live robot motion safety planning system based on multiple performance quantification constraints, such as Figure 1 and Figure 2 As shown, it includes an object detection module, which includes a visual sensor and a lidar sensor. The visual sensor is used to collect RGB image information in front of the robot, and the lidar sensor is used to collect three-dimensional point cloud data in front of the robot. The surface coordinate set A1 of the target object in front of the robot and the surface coordinate set A2 of the obstacle object are determined based on the RGB image information and the three-dimensional point cloud data.

[0033] Power companies use live-line robots to perform preparatory work on live grids, replacing manual labor. However, during these operations, both the target object and obstacles are in motion, making it difficult for traditional live-line robot arms to accurately handle moving objects.

[0034] To solve the above problems, an object detection module is set up in the robot. The object detection module receives RGB image information and 3D point cloud data, extracts the shape features of the target object and obstacle objects in the RGB image information, and performs contour analysis based on the shape features to determine the target object and obstacle objects.

[0035] Based on the three-dimensional point cloud data, the geometric dimensions and spatial positioning of the target object and obstacle objects are constructed, and a three-dimensional scene model is constructed through spatial registration and coordinate transformation algorithms. A three-dimensional Cartesian coordinate system with the robot body as the coordinate origin is established; based on the recognition results of the RGB image information, the surface coordinate set A1 of the target object and the surface coordinate set A2 of the obstacle object are obtained from the three-dimensional Cartesian coordinate system to accurately represent the spatial positions of the target object and the obstacle objects.

[0036] The global planning module is connected to the object detection module data to receive the surface coordinate set A1 of the target object before the robot's mechanical arm moves, the surface coordinate set A2 of the obstacle object, and the initial coordinates of the end of the mechanical arm. Destination coordinates of the end of the robotic arm and safety distance d safe ,The global motion path of the end of the robot arm is planned through the global path generation method.

[0037] Determine the initial coordinates of the end of the robotic arm based on the parameter information of the robotic arm before the robot moves Determine the target coordinates of the end of the robotic arm based on the surface coordinate set A1 of the target object An embodiment of the present application is: target coordinates The safe distance d between the end of the robotic arm and the obstacle or target object is set manually based on work experience. safe .

[0038] The global path generation method includes: generating sampling points, obtaining the surface coordinate set A1 of the target object, the surface coordinate set A2 of the obstacle object, obtaining the initial coordinates of the end of the mechanical arm Destination coordinates and the safety distance is d safe ,by and A rectangular parallelepiped sampling space is constructed for the diagonal endpoints, and sampling points are uniformly arranged at a fixed density within the sampling space to generate a sampling point set S1. In one embodiment of the present application, the sampling points are distributed in a three-dimensional matrix, and the distribution density of the sampling points is manually set; therefore, the larger the volume of the sampling space, the greater the number of sampling points, but the distribution density remains unchanged.

[0039] According to the surface coordinate set A1 of the target object, the surface coordinate set A2 of the obstacle object and the safety distance d safe , filter out the safe distance d between the target object surface and the obstacle surface safe Several sampling points other than the above are defined as valid sampling points, and all valid sampling points together constitute a first valid sampling point set S2;

[0040] Generate the global motion path, set the number of adjacent valid sampling points K based on experience, and the starting point Or the temporary starting point p is connected to the K adjacent valid sampling points in the first valid sampling point set S2 to generate K global neighborhood graphs. Or the temporary starting point p can be connected to the adjacent valid sampling points to form an edge, and collision detection is performed on the edges connected in the K global neighborhood graphs, and only the k neighborhood graphs (k≤K) with no collision are retained. Determine whether the k adjacent valid sampling points include the end point Yes, indicating that the corresponding global neighborhood graph has reached the end and the fitting includes the end point The global neighborhood graph is used to generate the global motion path. If not, it means that the global neighborhood graph is not completed yet, and jump to the step of determining the adjacent valid sampling points.

[0041] Determine the adjacent valid sampling points and calculate the function value F of k neighborhood graphs respectively through the first function (O) , filter out the function value F (O)The smallest neighborhood graph is obtained, and the last valid sampling point of the global neighborhood graph is defined as the temporary starting point p, and the process jumps to the step of generating the global neighborhood graph.

[0042] The first function is:

[0043]

[0044] in, Indicates the initial coordinates of the end of the robotic arm The Euclidean distance to all valid sampling points in the connected global neighborhood graph, Represents the coordinates of all valid sampling points to target objects in the connected global neighborhood graph The Euclidean distance of .

[0045] During the global path generation process, the steps of generating the global motion path and determining the adjacent valid sampling points are executed cyclically, and the adjacent valid sampling points to be connected are determined in sequence until the destination coordinates are reached. Complete the connection of all valid sampling points in the global neighborhood graph and end the loop.

[0046] The local reconstruction module is used to receive the surface coordinate set A1 of the target object, the surface coordinate set A2 of the obstacle object, the starting point near the target object or the obstacle object during the operation of the robot's manipulator and end point The local path near the target object or obstacle in the motion path that has not been passed is updated according to the local path generation method.

[0047] The local path generation method includes: generating sampling points, during the operation of the robot's end arm, the local reconstruction module continuously receives the surface coordinate set A1 of the target object and the surface coordinate set A2 of the obstacle object, and defines the first valid sampling point of the global motion path near the target object or the obstacle object as the starting point The last valid sampling point is defined as the end point One embodiment of the present application is to artificially set a proximity distance D, and define the space within the proximity distance D on the surface of the target object or obstacle as the vicinity of the target object or obstacle.

[0048] Based on the surface coordinate set A1 of the target object, the surface coordinate set A2 of the obstacle object, the first sampling point set S1 and the safety distance d safe Screen out the safe distance d between the target object surface and the obstacle surface safe Several valid sampling points other than , and generate a second valid sampling point set S3;

[0049] Generate local path, starting point Or the temporary starting point t is connected to a neighboring valid sampling point in the second valid sampling point set S3 to generate a local area graph, and determines whether the neighboring valid sampling point is the end point Yes, indicating that the end of the local path has been reached and the fit includes the end point The local neighborhood graph is used to generate a local path. If not, jump to the step of determining the adjacent valid sampling points.

[0050] Determine the adjacent valid sampling points and calculate the starting points respectively through the second function F between adjacent valid sampling points, temporary starting point t and adjacent valid sampling points (k) Difference, judge the two F (k) Is the difference less than the convergence threshold ε? If yes, it means that the connecting edge has no collision and is passable, connecting the adjacent valid sampling points and the starting point Or adjacent valid sampling points and temporary starting point t, and define the adjacent valid sampling points as temporary starting point t, jump to the step of generating local path; No, it means that the connecting edge may collide, and the adjacent valid sampling points and starting point No connection, jump to the step of generating a local path. One embodiment of the present application is: the staff is based on the safety distance d safe Set the convergence threshold ε. The smaller the convergence threshold ε is, the closer the adjacent valid sampling points and the starting point are to the required Or the side of the temporary starting point t is shorter, when the side is far less than the safety distance d safe , the edge and the object must not collide.

[0051] The second function is:

[0052]

[0053] in, Indicates the starting point The Euclidean distance to the adjacent valid sampling point or temporary starting point t, Indicates the adjacent valid sampling point or temporary starting point t to the end point The computational complexity of the second function is lower than that of the first function, and it can quickly determine and connect adjacent valid sampling points, thereby improving the speed of local path generation.

[0054] The local path generated by the local path generation method only determines the relationship between two adjacent valid sampling points. This can only ensure that the local path can bypass obstacles or target objects without collision, but does not guarantee the brevity of the local path. To improve the brevity of the local path, a multi-constraint rule optimization unit is set up in the local reconstruction module to optimize the local path based on the safety distance and the minimum rotation angle of the robot arm, remove unnecessary valid sampling points in the local path, and reduce the probability of the local path taking a detour. The multi-constraint rule optimization unit includes the evaluation function:

[0055] Ji =(max(0,d safe -d i )*max(0,θ i -θ min ))

[0056] Among them, θ min is the minimum movement angle of the robot arm, θ i is the motion angle of the manipulator between the first i-th valid sampling point i in the local neighborhood graph and the i-1th valid sampling point in the local neighborhood graph, d i Represents the distance between the i-th valid sampling point in the local neighborhood graph and the target object, and J in the local neighborhood graph i The i-th valid sampling point with =0 is removed, and the remaining valid sampling points in the local neighborhood graph are refitted to generate the optimized local path.

[0057] The local path generation method and the multi-constraint rule optimization unit can run synchronously, realizing the generation and optimization of the local path at the same time. Even if the local path has not been completely generated, the end of the robot arm can move along the unfinished path, thereby improving the continuity of the movement of the end of the robot arm.

[0058] The safety verification module is used to continuously receive the surface coordinate set A1 of the target object, the surface coordinate set A2 of the obstacle object, the current coordinate P of the end of the robot arm and the local path during the operation of the robot arm, so as to perform safety verification on the local path until the robot arm moves to the target coordinate

[0059] After the local path within the global motion path is updated, the robot's end arm continues to move along the global motion path that includes several local paths. During the movement, the safety verification module continues to receive the surface coordinate set A1 of the target object, the surface coordinate set A2 of the obstacle object, the current coordinate P of the end arm, and the local path. Based on the current coordinate P of the end arm, the valid sampling points that have not passed through the global motion path are determined. If the number of valid sampling points that have not passed through is 0, the robot arm ends the movement.

[0060] No, based on the surface coordinate set A1 of the target object and the surface coordinate set A2 of the obstacle object, determine whether there is a valid sampling point whose distance from the surface of the obstacle object (or target object) is less than the safety distance d safe If yes, the local reconstruction module re-updates the corresponding local path. If not, the safety verification module continues to receive the next set of surface coordinate sets A1 of the target object, the surface coordinate set A2 of the obstacle object, the current coordinate P of the end of the robot arm and the updated local path, and re-safety-checks all the local paths that have not been passed.

[0061] In this way, the robot's global motion path is initially planned before the end arm moves. As the end arm moves, the local path is continuously optimized based on surface changes near the object until the end arm reaches its destination. This improves the robot's ability to handle dynamically moving objects, effectively resolving the problem of existing powered robots whose arms are unable to flexibly handle the movement of dynamically moving objects.

[0062] The embodiments of the present invention are provided for the purpose of illustration and description. Although the embodiments of the present invention have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limitations of the present invention. Ordinary technicians in this field can change, modify, replace and modify the above embodiments within the scope of the present invention.

Claims

1. A live robot motion safety planning system based on multiple performance quantification constraints, characterized by: include: The object detection module is used to collect RGB image information and 3D point cloud data, and determine the surface coordinate set A1 of the target object in front of the robot and the surface coordinate set A2 of the obstacle object based on the RGB image information and 3D point cloud data; The global planning module is used to receive the surface coordinate set A1 of the target object before the robot's manipulator moves, the surface coordinate set A2 of the obstacle object, and the initial coordinates of the end of the manipulator. Destination coordinates of the end of the robotic arm and safety distance d safe , and then plan the global motion path of the end of the robot's manipulator through the global path generation method; The local reconstruction module is used to receive the surface coordinate set A1 of the target object, the surface coordinate set A2 of the obstacle object, the starting point near the target object or the obstacle object during the operation of the robot's manipulator and end point Update the local path near the target object or obstacle in the motion path that has not been passed according to the local path generation method; The safety verification module is used to continuously receive the surface coordinate set A1 of the target object, the surface coordinate set A2 of the obstacle object, the current coordinate P of the end of the robot arm and the local path during the operation of the robot arm, so as to perform safety verification on all local paths until the robot arm moves to the target coordinate 2. A live robot motion safety planning system based on multiple performance quantification constraints as claimed in claim 1, characterized in that: The object detection module extracts shape features based on RGB image information, performs contour analysis based on the shape features to determine the target object and obstacle objects, uses three-dimensional point cloud data to construct the geometric dimensions and spatial positioning of the target object and obstacle objects, and obtains the surface coordinate set A1 of the target object and the surface coordinate set A2 of the obstacle objects based on the results of the contour analysis.

3. The live robot motion safety planning system based on multiple performance quantification constraints as claimed in claim 1, characterized in that: The global path generation method comprises: Generate the first valid sampling point set S2 and obtain the initial coordinates of the end of the mechanical arm Destination coordinates and the safety distance is d safe ,by and Construct a rectangular parallelepiped sampling space for the diagonal endpoints, uniformly distribute the sampling points with a fixed density in the sampling space to generate a sampling point set S1, and screen out the safe distance d between the target object surface and the obstacle surface safe Several valid sampling points other than , and generate a first valid sampling point set S2; Generate global motion path, starting point Get the temporary starting point p and connect the K adjacent valid sampling points in the first valid sampling point set S2 to generate K global neighborhood graphs. Perform collision detection on the edges connected by the K global neighborhood graphs and only retain the k global neighborhood graphs with no collision edges. Determine whether the k global neighborhood graphs contain the end point. Yes, the fit includes the endpoint The global neighborhood graph is used to generate the global motion path. If not, jump to the step of determining the adjacent valid sampling points. Determine the global effective sampling points and calculate the function values ​​F of k global neighborhood graphs respectively through the first function (O) , the function value F (O) The last valid sampling point of the minimum global neighborhood graph is defined as the temporary starting point p, and the process jumps to the step of generating the global neighborhood graph.

4. The integrated live robot motion safety planning system based on multi-performance quantification and constraints as described in claim 3 is characterized in that: The first function is: in, Indicates the initial coordinates of the end of the robotic arm The Euclidean distance to all valid sampling points in the connected global neighborhood graph, Represents the coordinates of all valid sampling points to target objects in the connected global neighborhood graph The Euclidean distance of .

5. The integrated live robot motion safety planning system based on multi-performance quantification and constraints as claimed in claim 1, characterized in that: The local path generation method comprises: Generate a second valid sampling point set S3, receive the surface coordinate set A1 of the target object and the surface coordinate set A2 of the obstacle object, and define the first valid sampling point of the global motion path near the target object or obstacle as the starting point The last valid sampling point is defined as the end point Based on the surface coordinate set A1 of the target object, the surface coordinate set A2 of the obstacle object, the first sampling point set S1 and the safety distance d safe Screen out the safe distance d between the target object surface and the obstacle surface safe Several valid sampling points other than , and generate a second valid sampling point set S3; Generate local path, starting point Or the temporary starting point t is connected to a neighboring valid sampling point in the second valid sampling point set S3 to generate a local area graph, and it is determined whether the neighboring valid sampling point is the end point Yes, the fit includes the endpoint The local neighborhood graph is used to generate a local path. If not, jump to the step of determining the adjacent valid sampling points. Determine the local effective sampling points and calculate the starting points respectively through the second function F between adjacent valid sampling points, temporary starting point t and adjacent valid sampling points (k) Difference, judge the two F (k) Is the difference less than the convergence threshold ε? If yes, connect the adjacent valid sampling points to the starting point Or connect the adjacent valid sampling points with the temporary starting point t, and define the adjacent valid sampling points as the new starting point t, and jump to the step of generating a local path; otherwise, the adjacent valid sampling points are not connected to the starting point Or a temporary starting point t connection to jump to the step of generating a local path.

6. The integrated live robot motion safety planning system based on multi-performance quantification and constraints as described in claim 5, characterized in that: The second function is: in, Indicates the starting point The Euclidean distance to the adjacent valid sampling point or temporary starting point t, Indicates the adjacent valid sampling point or temporary starting point t to the end point The Euclidean distance of .

7. The integrated live robot motion safety planning system based on multi-performance quantification and constraints as claimed in claim 5, characterized in that: The local reconstruction module includes a multi-constraint rule optimization unit for correcting the local path, including an evaluation function: J i =(max(0,d safe -d i )*max(0,θ i -θ min )) Among them, θ min is the minimum movement angle of the robot arm, θ i is the motion angle of the manipulator between the first i-th valid sampling point i and the i-1th valid sampling point, d i Represents the distance between the i-th valid sampling point and the target object, and J i The i-th valid sampling point with =0 is eliminated, and the remaining valid sampling points are refitted to optimize the local path.

8. The integrated live robot motion safety planning system based on multi-performance quantification and constraints as claimed in claim 5, characterized in that: The safety verification module includes: determining the valid sampling points that have not passed the global motion path according to the current coordinate P of the end of the manipulator, judging whether the number of valid sampling points that have not passed is 0, if yes, the manipulator ends the movement, if not, judging whether there are any valid sampling points that have not passed the distance from the obstacle or the target object surface that is less than the safety distance d safe If yes, the local reconstruction module re-updates the corresponding local path. If no, it continues to receive the next set of surface coordinate sets A1 of the target object, the surface coordinate set A2 of the obstacle object, the current coordinates P of the end of the robotic arm, and the local path.

Images