CN-121733589-B - Manipulator cross-mode sensing control system for chemical emergency treatment

CN121733589BCN 121733589 BCN121733589 BCN 121733589BCN-121733589-B

Abstract

The invention relates to the technical field of mechanical arms, in particular to a cross-mode sensing control system of a mechanical arm for chemical emergency treatment, which comprises an environment scanning module, a visual positioning module, an obstacle avoidance planning module, a pose compensation module, a decision mapping module and a grabbing updating module, wherein the operation area of the mechanical arm is pre-scanned to obtain environment state information; the method comprises the steps of acquiring an operation view angle image of a manipulator, resolving visual features of the operation view angle image to obtain pose information, carrying out obstacle avoidance correction on a motion track of the manipulator to obtain a collision-free approach path, fusing the operation view angle image and multi-mode micro-touch signals, carrying out error compensation on the pose information to obtain an accurate grabbing pose, mapping physical properties of the object surface perceived by the multi-mode micro-touch signals to obtain an adaptive grabbing instruction, controlling the manipulator to grab by the adaptive grabbing instruction, and updating environment state information.

Inventors

XIAO WENXIA
ZHANG HAIYING

Assignees

呼伦贝尔职业技术学院

Dates

Publication Date: 20260508
Application Date: 20260228

Claims (8)

1. The utility model provides a manipulator cross-mode perception control system towards emergent handling of chemical industry, its characterized in that, the system includes environment scanning module, vision positioning module, keeps away barrier planning module, position appearance compensation module, decision mapping module and snatchs update module, wherein: The environment scanning module is used for carrying out three-dimensional environment pre-scanning on an operation area of the manipulator to obtain environment state information of the manipulator; The visual positioning module is used for acquiring an operation view angle image of the manipulator based on the environmental state information, and resolving visual characteristics of the operation view angle image to obtain pose information of the manipulator; the obstacle avoidance planning module is used for carrying out obstacle avoidance correction on the motion trail of the manipulator based on the pose information and the environmental state information to obtain a collision-free approach path of the manipulator; The pose compensation module is used for fusing the operation visual angle image and the multi-mode micro-touch signals acquired by the manipulator based on the collision-free approach path so as to perform error compensation on the pose information and obtain the accurate grabbing pose of the manipulator; The decision mapping module is configured to perform decision mapping on physical properties of the object surface perceived by the multi-mode micro-touch signal based on the accurate grabbing pose, to obtain an adaptive grabbing instruction of the manipulator, where the decision mapping module includes: determining an expected contact area between the manipulator and the target object according to the accurate grabbing pose; extracting contact force distribution information and surface roughness information related to the expected contact area from the multi-mode micro-touch signal to obtain the surface physical attribute of the target object; performing bidirectional matching on the surface physical attribute and a grabbing parameter in a preset grabbing strategy library to obtain grabbing strength and grabbing speed of the manipulator, wherein the bidirectional matching comprises the following steps: screening out candidate grabbing strength intervals and candidate grabbing speed intervals in the preset grabbing strategy library according to the surface physical attribute; selecting initial grabbing force from the candidate grabbing force interval to control the manipulator to apply pre-contact pressure to the surface of the target object, and synchronously monitoring the change rate of the contact force in the multi-mode micro-touch signal; Based on the change rate of the contact force, judging the deformation state of the target object under the initial grabbing force to obtain deformation response characteristics of the target object; Determining the initial grabbing speed of the candidate grabbing speed interval according to the deformation response characteristics; According to the initial grabbing force and the initial grabbing speed, controlling the manipulator to perform a trial grabbing action on the target object, and collecting a sliding signal in the trial grabbing process in real time; Based on the intensity of the sliding signal, synchronously correcting the initial grabbing force and the initial grabbing speed to obtain the grabbing force and the grabbing speed of the manipulator; encoding the grabbing force and the grabbing speed into an adaptive grabbing instruction of the manipulator; the grabbing updating module is used for controlling the manipulator to grab according to the self-adaptive grabbing instruction and updating the environment state information according to the grabbing state.
2. The manipulator cross-mode sensing control system for chemical emergency treatment according to claim 1, wherein the environment scanning module is specifically configured to, when executing three-dimensional environment pre-scanning on a working area of a manipulator to obtain environment state information of the manipulator: Controlling a three-dimensional laser scanner carried by a manipulator to scan an operation area of the manipulator to obtain original point cloud data of the operation area; Performing noise point elimination processing on the original point cloud data to obtain denoised point cloud data of the operation area; performing three-dimensional reconstruction on the denoised point cloud data to obtain a three-dimensional point cloud distribution map of the operation area; and taking the three-dimensional point cloud distribution map as the environmental state information of the manipulator.
3. The manipulator cross-mode sensing control system for chemical emergency treatment according to claim 1, wherein the visual positioning module is configured to, when executing the operation view angle image based on the environmental state information, collect the operation view angle image of the manipulator, and calculate the visual characteristics of the operation view angle image to obtain pose information of the manipulator: According to the environmental state information, adjusting the acquisition angle of a vision sensor carried by the manipulator, so that the field of view of the vision sensor covers the operation area to acquire an operation visual angle image of the manipulator; Gray level transformation is carried out on the operation view angle image, and binarization processing is carried out on the transformed image, so that a binarization image of the manipulator is obtained; Performing edge detection on the binarized image to obtain a contour feature image of the manipulator; and carrying out pose calculation on the spatial position and the pose angle of the manipulator based on the outline characteristic image to obtain pose information of the manipulator.
4. The chemical emergency treatment-oriented manipulator cross-mode perception control system according to claim 1, wherein the obstacle avoidance planning module is configured to, when executing obstacle avoidance correction on a motion track of the manipulator based on the pose information and the environmental state information to obtain a collision-free approach path of the manipulator: Performing track planning on the current pose information of the manipulator and the target grabbing position based on the pose information to obtain an initial motion track of the manipulator; According to the distribution positions of the obstacles in the environmental state information, carrying out sectional detection on the initial motion track to obtain a collision risk track section of the manipulator; selecting avoidance intermediate points on the periphery of the obstacle according to the collision risk track section; performing obstacle avoidance correction on the collision risk track section based on the avoidance intermediate point to obtain a corrected track section of the manipulator; And sequentially connecting the non-collision risk section in the initial motion track with the correction track section to obtain a collision-free approach path of the manipulator.
5. The chemical emergency treatment-oriented manipulator cross-modal awareness control system of claim 1, wherein the pose compensation module, when executing the collision-free approach path based, is specifically configured to, when fusing the operation view angle image with the multi-modal micro-haptic signals acquired by the manipulator: controlling the manipulator to move to a target object along the collision-free approach path; Collecting a contact force signal and a vibration signal of the target object by using a miniature touch sensor array integrated by the fingertip at the tail end of the manipulator to obtain a multi-mode micro touch signal of the target object; and performing time sequence alignment on the operation visual angle image and the multi-mode micro-touch signals to obtain the visual-touch perception data of the manipulator.
6. The chemical emergency treatment-oriented manipulator cross-mode sensing control system according to claim 5, wherein the pose compensation module is specifically configured to, when performing error compensation on the pose information to obtain an accurate grabbing pose of the manipulator: Extracting the edge contour of the target object from the operation view angle image to obtain the visual contour characteristic of the target object; Extracting contact point distribution and local geometric characteristics of the surface of the target object from the multi-mode micro-touch signal to obtain the touch geometric characteristics of the target object; Performing space reconstruction on the visual outline features and the tactile geometric features to obtain a fusion feature map of the manipulator; And carrying out residual error compensation on the pose information based on the fusion feature map to obtain the accurate grabbing pose of the manipulator.
7. The chemical emergency treatment-oriented manipulator cross-mode perception control system according to claim 6, wherein the pose compensation module is configured to, when performing residual compensation on the pose information based on the fusion feature map to obtain an accurate grabbing pose of the manipulator, specifically: extracting local feature points of the target object according to the fusion feature map to construct an actual measurement feature point set of the target object; determining a theoretical feature point set corresponding to the actual measurement feature point set of the target object under the theoretical pose according to the pose information; Performing differential analysis on the actual measurement characteristic point set and the theoretical characteristic point set to obtain a residual vector group of the manipulator; and carrying out weighted average on residual vectors in the residual vector group to obtain a comprehensive residual vector of the manipulator, wherein the calculation formula of the comprehensive residual vector is as follows: ; in the formula, For the said integrated residual vector(s), For the number of feature points in the set of measured feature points, Is the first The preset weight coefficients of the individual feature points, Centralizing the measured characteristic points The position coordinates of the individual measured feature points, Concentration of the theoretical feature points Position coordinates of the theoretical feature points; and based on the comprehensive residual vector, optimizing and compensating the pose information to obtain the accurate grabbing pose of the manipulator.
8. The chemical emergency treatment-oriented manipulator cross-mode sensing control system according to claim 1, wherein the grabbing update module is configured to, when executing the control of the manipulator to grab with the adaptive grabbing instruction, update the environmental state information according to a grabbing state: the self-adaptive grabbing instruction is sent to an execution terminal of the manipulator, and the manipulator is controlled to grab the target object; In the grabbing operation process, monitoring the contact state between the manipulator and the target object and the position change of the target object in real time to obtain grabbing state information of the manipulator; according to the grabbing state information, grabbing the target object is judged; when the grabbing is successful, removing the target object from the original position in the environment state information; And when the grabbing is failed, updating the position information of the target object in the environment state information according to the current position of the target object.

Description

Manipulator cross-mode sensing control system for chemical emergency treatment Technical Field The invention relates to the technical field of mechanical arms, in particular to a cross-mode sensing control system of a mechanical arm for chemical emergency treatment. Background Under the emergent scene of handling of chemical industry, the manipulator is as remote operation core equipment, manipulator perception and control performance directly decides the efficiency and the security of handling the operation, current manipulator control system is based on single vision perception to realize gesture recognition and track planning in many cases, in complicated chemical industry operation environment, only rely on visual information to be difficult to accurately catch the three-dimensional space characteristics of operation region, there is obvious limitation to the discernment of barrier and avoid, easily lead to motion track planning deviation, can't form stable crashless operation route, restricted the operation suitability of manipulator in the emergent scene of chemical industry. Meanwhile, the traditional manipulator lacks a multi-mode sensing fusion and dynamic pose compensation mechanism, the self-adaptive adjustment of the grabbing strategy cannot be realized by combining physical attributes of the object surface in the grabbing process, the grabbing action is finished only through preset parameters, the problems of pose error accumulation and improper grabbing force and speed matching are easily caused, the grabbing success rate is reduced, secondary safety risks such as chemical material leakage and equipment collision are possibly caused by improper operation, environmental state information cannot be updated in real time in the operation process, the operation requirement of environmental dynamic change in the chemical emergency treatment is difficult to adapt, and the actual technical requirements of the chemical emergency treatment cannot be met by the overall control precision and the operation intellectualization level. Disclosure of Invention In order to achieve the above purpose, the invention provides a manipulator cross-mode perception control system for chemical emergency treatment, which is characterized by comprising an environment scanning module, a visual positioning module, an obstacle avoidance planning module, a pose compensation module, a decision mapping module and a grabbing updating module, wherein: The environment scanning module is used for carrying out three-dimensional environment pre-scanning on an operation area of the manipulator to obtain environment state information of the manipulator; The visual positioning module is used for acquiring an operation view angle image of the manipulator based on the environmental state information, and resolving visual characteristics of the operation view angle image to obtain pose information of the manipulator; the obstacle avoidance planning module is used for carrying out obstacle avoidance correction on the motion trail of the manipulator based on the pose information and the environmental state information to obtain a collision-free approach path of the manipulator; The pose compensation module is used for fusing the operation visual angle image and the multi-mode micro-touch signals acquired by the manipulator based on the collision-free approach path so as to perform error compensation on the pose information and obtain the accurate grabbing pose of the manipulator; the decision mapping module is used for performing decision mapping on physical properties of the object surface perceived by the multi-mode micro-touch signal based on the accurate grabbing pose to obtain an adaptive grabbing instruction of the manipulator; the grabbing updating module is used for controlling the manipulator to grab according to the self-adaptive grabbing instruction and updating the environment state information according to the grabbing state. In a preferred embodiment, the environment scanning module is specifically configured to, when performing three-dimensional environment pre-scanning on a working area of a manipulator to obtain environment state information of the manipulator: Controlling a three-dimensional laser scanner carried by a manipulator to scan an operation area of the manipulator to obtain original point cloud data of the operation area; Performing noise point elimination processing on the original point cloud data to obtain denoised point cloud data of the operation area; performing three-dimensional reconstruction on the denoised point cloud data to obtain a three-dimensional point cloud distribution map of the operation area; and taking the three-dimensional point cloud distribution map as the environmental state information of the manipulator. In a preferred embodiment, the visual positioning module is specifically configured to, when executing the operation view angle image of the manipulator based on the environmental statu