Robotic Arm Remote Control Teaching System

Abstract

The present invention discloses a remote control teaching system for a robotic arm, which relates to the field of virtual reality technology and solves the problems of lack of systematic analysis of path features, which easily leads to redundancy and mutation in the motion trajectory of the robotic arm. The present invention innovatively locks the mutation inflection point through inflection point angle analysis at the path feature verification end, effectively identifies abnormal turning points in the path, and lays the foundation for subsequent optimization; the preliminary optimization end and the secondary optimization end adopt a hierarchical optimization strategy, first generating a preliminary optimization path by connecting the mutation inflection points to reduce invalid movement of the robotic arm, and then iteratively screening the shortest interference-free path to achieve deep optimization of the path, significantly improving the efficiency and safety of path planning; the execution end accurately controls the robotic arm according to the optimized execution path, which not only reduces the energy consumption and time cost of the robotic arm operation, but also effectively avoids collision risks through a scientific path optimization mechanism.

Description

Technical Field

[0001] The present invention relates to the field of virtual reality (VR) technology, and in particular to a robotic arm remote control teaching system. Background Art

[0002] In the fields of industrial automation, hazardous environment operations, etc., the application of robotic arms is becoming more and more extensive, and the accuracy and efficiency of their path planning and control have become key technical points; traditional robotic arm teaching methods mostly rely on manual on-site operations or preset fixed programs, and have problems such as poor flexibility, low efficiency, and difficulty in adapting to complex dynamic environments.

[0003] With the development of remote control technology, remote teaching systems can break through spatial limitations, but existing solutions still face many challenges: First, path planning is mostly based on experience or simple algorithms, lacking a systematic analysis of path characteristics, which can easily lead to redundancy and mutation in the robot arm's motion trajectory, reducing operational efficiency and increasing energy consumption; second, in complex environments, path planning is difficult to effectively avoid obstacles, which may cause collision risks and threaten equipment safety; third, existing optimization strategies are single and cannot meet the diverse needs of path optimization under different working conditions.

[0004] Therefore, there is an urgent need for a remote control teaching system that can accurately analyze path characteristics, efficiently optimize motion trajectories, and ensure the safe and stable operation of the robotic arm. Summary of the Invention

[0005] In response to the shortcomings of the existing technology, the present invention provides a remote control teaching system for a robotic arm, which solves the problems of lack of systematic analysis of path characteristics, which easily leads to redundancy and mutation in the robotic arm's motion trajectory.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: a remote control teaching system for a robotic arm, comprising:

[0007] The virtual scene construction end constructs a virtual scene of the environment in which the robotic arm works and generates a virtual scene;

[0008] On the virtual path generation side, operators conduct actual operations in the constructed virtual scene to confirm the virtual path;

[0009] The path feature verification end performs feature verification on the recorded virtual path, calibrates the path inflection points existing in the virtual path, and locks the sudden change inflection point from the calibrated path inflection points. The specific method is as follows:

[0010] The path inflection points in the virtual path are calibrated one by one: the different path points associated with the virtual path are recorded as pending points, and the previous and next sets of path points adjacent to the pending point are determined. Based on the pending point and the two adjacent sets of points, the characteristic angle belonging to the pending point is determined. If the characteristic angle = 180°, no calibration is performed. If the characteristic angle ≠ 180°, the pending point is calibrated as a path inflection point.

[0011] Confirm the sudden inflection point from the marked path inflection points: Based on the marked path inflection point, the path point before the path inflection point is recorded as the front point, the path point after the path inflection point is recorded as the back point, and the part of the line segment between the front point and the path inflection point is recorded as the front feature segment. The front feature segment is extended according to the moving direction of the front point toward the path inflection point to confirm a set of extended segments. The part of the segment between the back point and the path inflection point is recorded as the back feature segment, and the angle between the extended segment and the back feature segment is recorded as JA i , where i represents different path inflection points, if JA i ≤Y1, no processing is performed. If JA i >Y1, the associated path inflection point is marked as a sudden inflection point, and its Y1 is the preset value;

[0012] At the initial optimization end, based on the multiple mutation inflection points marked in the virtual path and the actual spatial positions, the mutation inflection points determined in sequence are connected to generate the first set of optimized paths belonging to the virtual path. The specific method is as follows:

[0013] Based on the multiple mutation inflection points marked in the virtual path, adjacent mutation inflection points are connected in sequence to confirm the preferred single path between the adjacent mutation inflection points and to confirm whether there is any interference with the preferred single path in the virtual scene. If so, the associated travel path of the adjacent mutation inflection points in the virtual path is determined and the determined travel path is recorded as the determined path. If not, the preferred single path is recorded as the determined path.

[0014] Connecting the plurality of groups of determined paths recorded in sequence to generate a first group of optimized paths belonging to the virtual path;

[0015] Identify the original path length of the virtual path and record it as L1. Then confirm the path length of the first set of optimized paths and record it as L2. Use: (L1-L2)÷L1=YH to determine the optimized value YH, and display the confirmed optimized value YH on the display terminal.

[0016] On the secondary optimization side, the first set of optimized paths is optimized. From the several groups of mutation inflection points marked in the first set of optimized paths, selected inflection points are locked. Then, the paths of the several groups of selected inflection points are confirmed. The execution path is locked from the confirmed groups of paths. The specific method is as follows:

[0017] S1. The first set of optimized paths determined is taken as the pending path, and several sudden inflection points existing in the pending path are recorded as pending points. The positions of the pending points appearing in sequence are recorded and labeled as k, where k = 1, 2, ..., n. When k is 1, it means that the pending point belongs to the initial point of the pending path, and when k is n, it means that the pending point belongs to the end point of the pending path;

[0018] S2. Prioritize the two groups of undetermined points with k=1 and k=n as selected inflection points, connect the two groups of selected inflection points, determine a set of characteristic paths, and identify whether there are interference objects in the virtual scene for this characteristic path. If there are interference objects, no calibration is performed and subsequent processing is performed. If not, this path is used as the execution path, and the execution path determination process is completed;

[0019] S3. When the first set of characteristic paths is determined, a set of sudden inflection points is selected from k∈[2, n-1] in sequence and recorded as pending points. The characteristic paths are adjusted according to the positions of the pending points to confirm the second set of characteristic paths. It is identified whether the confirmed sets of characteristic paths have interference objects in the virtual scene. If so, no calibration is performed. If not, the corresponding characteristic paths are recorded as candidate paths. If only one set of candidate paths exists in this processing process, this candidate path is used as the execution path. If multiple sets of candidate paths exist, the shortest path is selected from the multiple sets of candidate paths and the selected shortest path is used as the execution path. If no set of candidate paths exists in the multiple sets of characteristic paths, the subsequent processing process is executed again.

[0020] S4. Sequentially add unselected mutation inflection points to the multiple sets of characteristic paths associated with the previous set of processing processes, perform adjustments and confirmations to obtain subsequent characteristic paths, and use the same confirmation method as in S3 to lock the execution path or execute subsequent processing processes.

[0021] S5. Repeat S4 until several sudden inflection points are locked as selected inflection points, and then stop to confirm the execution path.

[0022] Preferably, the virtual scene construction end performs a three-dimensional scan of the environment around the robotic arm based on the camera and lidar sensor set on the robotic arm, confirms the position features associated with different pixels in the environment in sequence, and generates a virtual scene belonging to the corresponding robotic arm working environment based on the different pixel position features confirmed in sequence.

[0023] Preferably, during the generation process of the virtual path: a relevant operator wears a VR device to enter the virtual scene, and actually controls the virtual arm in the virtual scene, records the running trajectory generated by the virtual arm, and generates a virtual path.

[0024] Preferably, the execution end controls the operation of the associated robotic arm according to the determined execution path.

[0025] Preferably, each group of virtual paths is associated with a group of operating nodes, and the operating nodes of the virtual arm are consistent with the operating nodes of the robotic arm.

[0026] The present invention provides a remote control teaching system for a robotic arm. Compared with the prior art, it has the following advantages:

[0027] This invention innovatively uses inflection point angle analysis to identify sudden inflection points at the path feature verification end, effectively identifying abnormal turns in the path and laying the foundation for subsequent optimization. The preliminary optimization end and the secondary optimization end adopt a hierarchical optimization strategy. First, the preliminary optimization path is generated by connecting sudden inflection points to reduce invalid motion of the robot arm. Then, the shortest interference-free path is iteratively selected to achieve deep optimization of the path, significantly improving the efficiency and safety of path planning.

[0028] The execution end precisely controls the robotic arm based on the optimized execution path, ensuring efficient and stable operation in complex scenarios. The entire system not only reduces the energy consumption and time costs of robotic arm operations, but also effectively avoids collision risks through a scientific path optimization mechanism, greatly enhancing the system's reliability and practicality. It is suitable for a variety of applications, including industrial manufacturing and hazardous environment operations. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 This is a schematic diagram of the principle framework of the present invention;

[0030] Figure 2 Schematic diagram for determining the angle between the front and rear characteristic segments of the present invention. DETAILED DESCRIPTION

[0031] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0032] First embodiment:

[0033] See also Figure 1 The present application provides a remote control teaching system for a robotic arm, including a virtual scene construction end, a virtual path generation end, a path feature verification end, a preliminary optimization end, a secondary optimization end, and an execution end, wherein the virtual scene construction end, the virtual path generation end, the path feature verification end, the preliminary optimization end, the secondary optimization end, and the execution end are electrically connected from an output node to an input node in sequence:

[0034] Among them, the virtual scene construction end constructs a virtual scene of the environment in which the robotic arm works, and performs a three-dimensional scan of the environment around the robotic arm based on the camera and laser radar sensors set on the robotic arm, and confirms the position features associated with different pixels in the environment in turn, and generates a virtual scene belonging to the corresponding robotic arm working environment based on the different pixel position features confirmed in turn. Specifically, in the specific construction process of the virtual scene construction process of this part, based on the pixel features captured by the corresponding camera and the position features of the corresponding pixel features determined by the laser radar, the spatial features of the specified pixel can be confirmed, and then based on the different spatial features associated with different pixels, the virtual scene of the robotic arm working environment can be generated. Therefore, this part of the construction process is relatively common in the existing technology, so it will not be described in detail here;

[0035] Among them, at the virtual path generation end, the relevant operators perform practical operations in the constructed virtual scene by wearing VR equipment, and in the process of practical operation, confirm the operation trajectory of the corresponding virtual arm in the virtual scene, and record the confirmed operation trajectory as a virtual path. Different operation nodes have different virtual paths. This part of the virtual path is generated by the operator himself. The corresponding operator controls the corresponding virtual arm to perform relevant practical operations based on the constructed virtual scene. In the process of practical operation, the virtual path associated with each different node can be recorded, and the recorded virtual path is then transmitted to the control terminal of the robotic arm to control the robotic arm to operate at the same frequency, so as to ensure the actual operation process of the corresponding robotic arm and enable the robotic arm to adapt to a variety of different operation scenarios;

[0036] The path feature verification end performs feature verification on the recorded virtual path, calibrates the path inflection points existing in the virtual path, and locks the sudden inflection point from the calibrated path inflection points. The determined sudden inflection point is then calibrated in the virtual path. The sudden inflection point locking method is as follows:

[0037] The path inflection points in the virtual path are calibrated one by one: the different path points associated with the virtual path are recorded as pending points, and the previous and next sets of path points adjacent to the pending point are determined. Based on the pending point and the two adjacent sets of points, the characteristic angle belonging to the pending point is determined. If the characteristic angle = 180°, no calibration is performed. If the characteristic angle ≠ 180°, the pending point is calibrated as a path inflection point.

[0038] Combine Figure 2, confirm the sudden inflection point from the marked path inflection points: based on the marked path inflection point, record the path point before the path inflection point as the front point, record the path point after the path inflection point as the back point, record the part of the line segment between the front point and the path inflection point as the front feature segment, and extend the front feature segment according to the moving direction of the front point toward the path inflection point to confirm a group of extended segments, and then record the part of the segment between the back point and the path inflection point as the back feature segment, and record the angle between the extended segment and the back feature segment as JA i , where i represents different path inflection points, if JA i ≤Y1, no processing is performed. If JA i >Y1, the associated path inflection point is marked as a sudden inflection point, where Y1 is a preset value, the specific value of which is determined by the operator based on experience, and Y1 is generally set to 10°;

[0039] Specifically, after the virtual path is generated, there will be corresponding point changes. When the adjacent points have corner changes, the characteristic angle generated between the two groups of points changes, and the characteristic angle is no longer 180°, which means that the corresponding point belongs to the corresponding path inflection point. The inflection point angles of the path inflection points are then reconfirmed in turn, and the sudden inflection points are reconfirmed based on the reconfirmed inflection point angles. When the corresponding inflection point angle does not exceed the corresponding 10°, the inflection angle of the path is low, and there is no need to perform specific calibration of the sudden inflection points. If it exceeds the corresponding 10°, it is necessary to calibrate the sudden inflection points in the path, so as to confirm the path characteristics in the virtual path in turn.

[0040] In the preliminary optimization, based on the multiple mutation inflection points marked in the virtual path and the actual spatial positions of the mutation inflection points, the mutation inflection points determined in sequence are connected to generate the first set of optimized paths belonging to the virtual path. The specific method of generating the paths is as follows:

[0041] Based on several mutation inflection points marked in the virtual path, adjacent mutation inflection points are connected in sequence to confirm the preferred single path between the adjacent mutation inflection points, and to confirm whether the preferred single path has interference objects in the virtual scene (that is, the situation where the path and other objects have spatial feature intersections): If so, the behavior path associated with the adjacent mutation inflection points in the virtual path is determined, and the determined behavior path is recorded as the determined path; if not, the preferred single path is recorded as the determined path, and its interference object is the situation where the determined preferred single path and the object have spatial intersections, that is, the corresponding single path passes through objects in the virtual scene or other situations, which is a situation where there is an interference object;

[0042] Connecting the plurality of groups of determined paths recorded in sequence to generate a first group of optimized paths belonging to the virtual path;

[0043] Identify the original path length of the virtual path and record it as L1, then confirm the path length of the first set of optimized paths and record it as L2, use: (L1-L2)÷L1=YH to confirm the optimized value YH, and display the confirmed optimized value YH through the display terminal. Specifically, after the first set of optimized paths is optimized and determined, the path length of the robot arm must be lower than the overall line length of the original virtual path. This part of the optimization process is to prevent the robot arm from doing useless work during actual operation. This method is used for path optimization to determine the first set of optimized paths for the corresponding path.

[0044] Second embodiment:

[0045] During the specific implementation process, the main difference between this embodiment and the above embodiment is that this embodiment mainly performs secondary optimization on the optimized path to achieve a more optimized path selection process, so that the robot arm can achieve the optimal movement path confirmation effect and quickly execute the corresponding operation process;

[0046] The secondary optimization end performs path optimization processing on the first set of optimized paths after determination. From the several sets of sudden inflection points marked in the first set of optimized paths, selected inflection points are locked. Then, the paths of the several sets of selected inflection points are confirmed. From the several confirmed paths, the execution path is locked.

[0047] The specific method of locking the execution path is as follows:

[0048] S1. The first set of optimized paths determined is taken as the pending path, and several sudden inflection points existing in the pending path are recorded as pending points. The positions of the pending points appearing in sequence are recorded and labeled as k, where k = 1, 2, ..., n. When k is 1, it means that the pending point belongs to the initial point of the pending path, and when k is n, it means that the pending point belongs to the end point of the pending path;

[0049] S2. Prioritize the two groups of undetermined points with k=1 and k=n as selected inflection points, connect the two groups of selected inflection points, determine a set of characteristic paths, and identify whether there are interference objects in the virtual scene for this characteristic path. If there are interference objects, no calibration is performed and subsequent processing is performed. If not, this path is used as the execution path, and the execution path determination process is completed;

[0050] S3. When the first set of characteristic paths is determined, a set of mutation inflection points is selected from k∈[2, n-1] in sequence and recorded as pending points. The characteristic paths are adjusted according to the positions of the pending points to confirm the second set of characteristic paths (the behavior of the corresponding paths is specifically changed after the adjustment, and a set of mutation inflection points is added between k=1 and k=n). It is identified whether the confirmed sets of characteristic paths have interference objects in the virtual scene. If so, no calibration is performed. If not, the corresponding characteristic paths are recorded as candidate paths. If there is only one set of candidate paths in this processing process, this candidate path is used as the execution path. If there are multiple sets of candidate paths, the shortest path is selected from the multiple sets of candidate paths and the selected shortest path is used as the execution path. If there is no set of candidate paths in the multiple sets of characteristic paths, the subsequent processing process is executed again.

[0051] S4. Sequentially add unselected mutation inflection points to the multiple sets of characteristic paths associated with the previous set of processing processes, perform adjustments and confirmations to obtain subsequent characteristic paths, and use the same confirmation method as in S3 to lock the execution path or execute subsequent processing processes.

[0052] S5. Repeat S4 until several sudden inflection points are locked as selected inflection points, and then confirm the execution path (that is, when all sudden inflection points are locked as selected inflection points, the execution path determined in the second optimization process and the optimization path associated with the first optimization state belong to the same path).

[0053] Specifically, from the determined groups of mutation inflection points, the associated inflection points are selected in turn, and then according to the step-by-step selection process, the associated groups of paths can be confirmed in turn. Then, from the confirmed groups of paths, the related path with the shortest path length is selected as the corresponding execution path. Subsequently, according to the corresponding selection process, the corresponding execution path is quickly determined.

[0054] Among them, the execution end controls the operation of the associated robotic arm according to the determined execution path.

[0055] Some of the data in the above formulas are dimensionless and numerically calculated. Meanwhile, the contents not described in detail in this specification belong to the prior art known to those skilled in the art.

[0056] The above embodiments are only used to illustrate the technical method of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. The robot arm remote control teaching system is characterized by: include: The virtual scene construction end constructs a virtual scene of the environment in which the robotic arm works and generates a virtual scene; On the virtual path generation side, operators conduct actual operations in the constructed virtual scene to confirm the virtual path; The path feature verification end performs feature verification on the recorded virtual path, calibrates the path inflection points existing in the virtual path, and locks the sudden change inflection point from the calibrated path inflection points. The specific method is as follows: The path inflection points in the virtual path are calibrated one by one: the different path points associated with the virtual path are recorded as pending points, and the previous and next sets of path points adjacent to the pending point are determined. Based on the pending point and the two adjacent sets of points, the characteristic angle belonging to the pending point is determined. If the characteristic angle = 180°, no calibration is performed. If the characteristic angle ≠ 180°, the pending point is calibrated as a path inflection point. Confirm the sudden inflection point from the marked path inflection points: Based on the marked path inflection point, the path point before the path inflection point is recorded as the front point, the path point after the path inflection point is recorded as the back point, and the part of the line segment between the front point and the path inflection point is recorded as the front feature segment. The front feature segment is extended according to the moving direction of the front point toward the path inflection point to confirm a set of extended segments. The part of the segment between the back point and the path inflection point is recorded as the back feature segment, and the angle between the extended segment and the back feature segment is recorded as JA i , where i represents different path inflection points, if JA i ≤Y1, no processing is performed. If JA i >Y1, the associated path inflection point is marked as a sudden inflection point, and its Y1 is the preset value; At the initial optimization end, based on the multiple mutation inflection points marked in the virtual path and the actual spatial positions, the mutation inflection points determined in sequence are connected to generate the first set of optimized paths belonging to the virtual path; On the secondary optimization side, the first set of optimized paths is optimized. From the several sets of mutation inflection points marked in the first set of optimized paths, the selected inflection points are locked, and then the paths of the several sets of selected inflection points are confirmed. The execution path is locked from the confirmed several sets of paths.

2. The robotic arm remote control teaching system according to claim 1, characterized in that: The virtual scene construction end performs a three-dimensional scan of the environment around the robotic arm based on the camera and lidar sensor installed on the robotic arm, confirms the position features associated with different pixels in the environment in sequence, and generates a virtual scene belonging to the corresponding robotic arm working environment based on the different pixel position features confirmed in sequence.

3. The robotic arm remote control teaching system according to claim 1, characterized in that: During the generation process of the virtual path: the relevant operator wears a VR device to enter the virtual scene, and actually controls the virtual arm in the virtual scene, records the running trajectory generated by the virtual arm, and generates a virtual path.

4. The robotic arm remote control teaching system according to claim 1, characterized in that: The specific method of generating the first set of optimization paths at the preliminary optimization end is: Based on the multiple mutation inflection points marked in the virtual path, adjacent mutation inflection points are connected in sequence to confirm the preferred single path between the adjacent mutation inflection points and to confirm whether there is any interference with the preferred single path in the virtual scene. If so, the associated travel path of the adjacent mutation inflection points in the virtual path is determined and the determined travel path is recorded as the determined path. If not, the preferred single path is recorded as the determined path. Connecting the plurality of groups of determined paths recorded in sequence to generate a first group of optimized paths belonging to the virtual path; Identify the original path length of the virtual path and record it as L1, then confirm the path length of the first group of optimized paths and record it as L2, use: (L1-L2)÷L1=YH to confirm the optimized value YH, and display the confirmed optimized value YH through the display terminal.

5. The robotic arm remote control teaching system according to claim 1, characterized in that: The specific method of locking the execution path at the secondary optimization end is as follows: S1. The first set of optimized paths determined is taken as the pending path, and several sudden inflection points existing in the pending path are recorded as pending points. The positions of the pending points appearing in sequence are recorded and labeled as k, where k = 1, 2, ..., n. When k is 1, it means that the pending point belongs to the initial point of the pending path, and when k is n, it means that the pending point belongs to the end point of the pending path; S2. Prioritize the two groups of undetermined points with k=1 and k=n as selected inflection points, connect the two groups of selected inflection points, determine a set of characteristic paths, and identify whether there are interference objects in the virtual scene for this characteristic path. If there are interference objects, no calibration is performed and subsequent processing is performed. If not, this path is used as the execution path, and the execution path determination process is completed; S3. When the first set of characteristic paths is determined, a set of sudden inflection points is selected from k∈[2, n-1] in sequence and recorded as pending points. The characteristic paths are adjusted according to the positions of the pending points to confirm the second set of characteristic paths. It is identified whether the confirmed sets of characteristic paths have interference objects in the virtual scene. If so, no calibration is performed. If not, the corresponding characteristic paths are recorded as candidate paths. If only one set of candidate paths exists in this processing process, this candidate path is used as the execution path. If multiple sets of candidate paths exist, the shortest path is selected from the multiple sets of candidate paths and the selected shortest path is used as the execution path. If no set of candidate paths exists in the multiple sets of characteristic paths, the subsequent processing process is executed again. S4. Sequentially add unselected mutation inflection points to the multiple sets of characteristic paths associated with the previous set of processing processes, perform adjustments and confirmations to obtain subsequent characteristic paths, and use the same confirmation method as in S3 to lock the execution path or execute subsequent processing processes. S5. Repeat S4 until several sudden inflection points are locked as selected inflection points, and then stop to confirm the execution path.

6. The robotic arm remote control teaching system according to claim 5, characterized in that: Also includes: The execution end controls the operation of the associated robotic arm according to the determined execution path.

7. The robotic arm remote control teaching system according to claim 1, characterized in that: Each group of virtual paths is associated with a group of operating nodes, and the operating nodes of the virtual arm are consistent with the operating nodes of the robotic arm.

Images