CN-121973180-A - Industrial mechanical arm path optimization system based on electromagnetic scanning

CN121973180ACN 121973180 ACN121973180 ACN 121973180ACN-121973180-A

Abstract

The invention discloses an industrial mechanical arm path optimization system based on electromagnetic scanning, which comprises an image recognition module for acquiring target area information, wherein the mechanical arm performs electromagnetic scanning according to a positioning result and acquires electromagnetic data in a space. The system builds a spatial environment model fusing the image and the electromagnetic information, extracts barrier boundaries and physical constraints, and forms a dynamic convex constraint set. In the path planning process, a differential evolution algorithm is adopted to generate a variant path, and after each round of evolution, the path is mapped into a constraint range through a POCS algorithm, so that iterative convergence of the path under environmental constraint is realized. And finally, selecting an optimal path according to the adaptability index of the path, and controlling the mechanical arm to execute actions. The invention realizes the joint modeling and dynamic optimization control of path planning and environmental data.

Inventors

LIU WENTING
ZENG ZIMENG
Deng Yushi
Zuo Xinyue
FAN MIN

Assignees

重庆大学

Dates

Publication Date: 20260505
Application Date: 20251217

Claims (10)

1. The industrial mechanical arm path optimization system based on electromagnetic scanning is characterized by comprising an image recognition module, a scanning control module, an environment modeling module, a constraint generation module, a path constraint module and a path output module; The image recognition module acquires an image of a target area through an image acquisition device arranged on the industrial mechanical arm, and extracts image boundary information and space positioning data of the target area according to color characteristics; the scanning control module controls the industrial mechanical arm to move to a target area according to the space positioning data, and performs near-field electromagnetic sampling through an electromagnetic scanning device at the tail end of the industrial mechanical arm to acquire original electromagnetic data; the environment modeling module performs joint modeling on the image boundary information and the original electromagnetic data to construct a space environment model for path planning; The constraint generation module extracts obstacle information and physical boundary conditions related to path planning based on a space environment model, and forms a dynamically updated convex constraint set; the path evolution module generates a plurality of path initial solutions meeting convex constraint conditions in a path search variable space, and performs differential evolution operation on the path initial solutions to obtain a plurality of variant path candidate solutions; the path constraint module calls a POCS algorithm to demap variant path candidates into a convex constraint set to obtain a path set meeting environmental constraints; And the path output module selects an optimal path according to the adaptability indexes of the paths of the path set which are determined to meet the environmental constraint, and converts the optimal path into a control instruction to execute the path action of the mechanical arm.
2. The electromagnetic scanning-based industrial robot path optimization system according to claim 1, wherein the image recognition module comprises an image acquisition unit, an image conversion unit, a parameter setting unit, a mask generation unit, a boundary extraction unit, and a coordinate analysis unit; The image acquisition unit acquires color image frame data of a target area through a camera device arranged at the tail end of the industrial mechanical arm; the image conversion unit converts color image frame data into HSV color space images; The parameter setting unit stores a parameter set for color segmentation; the mask generating unit extracts pixel areas which accord with set color characteristics from the HSV color space image according to the parameter set and generates a mask image containing a target area; when a pixel region conforming to the set color characteristics is extracted, carrying out priori region verification, neighborhood characteristic verification and/or time sequence consistency detection on the pixel region, and deleting the misjudged pixel region; the prior area verification means that a target fixed appearance range is preset in advance, and a pixel area which exceeds the target fixed appearance range and accords with the set color characteristic is marked as a misjudgment pixel area; neighborhood feature verification means that texture/gradient features of adjacent pixels of a mask region are calculated, and if the periphery has no pre-stored target typical texture features, the pixel region is marked as a misjudgment pixel region; The time sequence consistency detection is to compare mask positions of adjacent frames, and if the position change of the pixel region does not accord with a preset motion rule, the pixel region is marked as a misjudgment pixel region; the boundary extraction unit analyzes the continuous pixel structure in the mask image and extracts image boundary information of a target area in the image; the coordinate analysis unit analyzes the spatial positioning data of the target area in the image space according to the image boundary information and the imaging parameters of the image acquisition device.
3. The electromagnetic scan based industrial robot path optimization system of claim 2, wherein the set of parameters for color segmentation includes upper and lower thresholds for identifying hue, saturation, and brightness of the target region, the upper and lower thresholds being determined from the sample image.
4. The electromagnetic scan based industrial robot path optimization system of claim 1, wherein the scan control module comprises a target input unit, a motion planning unit, a gesture control unit, a position confirmation unit, a scan trigger unit, and a data binding unit: the target input unit receives the space positioning data output by the image recognition module and converts the space positioning data into target pose data in a three-dimensional working space; the motion planning unit generates a motion path instruction sequence of the industrial mechanical arm according to the target pose data; The gesture control unit receives a motion path instruction sequence, controls all joints of the mechanical arm to cooperatively move, and adjusts the tail end of the electromagnetic scanning device to a gesture vertically oriented to a target area; The position confirmation unit acquires a real-time difference value between the current position and the target pose in the movement process of the mechanical arm, and updates a movement path according to the real-time difference value so that the mechanical arm reaches the target pose; when the gesture control unit completes gesture adjustment and the position confirmation unit determines that the target gesture is reached, the scanning triggering unit controls the electromagnetic scanning device to perform near-field sampling of electromagnetic data at a preset scanning height h, and records the spatial gesture corresponding to the sampling moment; and the data binding unit carries out corresponding binding on the original electromagnetic data acquired by the sampling triggering unit and the three-dimensional pose information provided by the target receiving unit.
5. The electromagnetic scanning-based industrial robot path optimization system according to claim 1, wherein the environment modeling module comprises an image data receiving unit, an electromagnetic data receiving unit, a coordinate registering unit, an attribute fusion unit, a grid construction unit, and a model generating unit; The image data receiving unit receives the image boundary information output by the image recognition module; The electromagnetic data receiving unit receives the original electromagnetic data which is output by the scanning control module and is bound with the three-dimensional pose information; The coordinate registration unit performs coordinate mapping on pixel coordinates and three-dimensional pose information contained in the image boundary information to generate a corresponding relation between an image space and an actual working space; the attribute fusion unit matches each image boundary area with the electromagnetic signal intensity and the electromagnetic gradient value corresponding to the three-dimensional pose information according to the space corresponding relation, and forms a multi-dimensional combined attribute set comprising coordinates, electromagnetic amplitude values, change gradients and image boundaries; the grid construction unit divides the working space into three-dimensional grid units, and injects physical information in the multi-dimensional joint attribute sets into the corresponding three-dimensional grid units according to the multi-dimensional joint attribute sets, wherein the physical information comprises coordinates, electromagnetic amplitude values, change gradients and image characteristics; The model generation unit takes the three-dimensional grids generated by the grid construction unit as nodes, establishes a space connection relation between each three-dimensional grid unit, and constructs a space environment model containing node coordinates, electromagnetic attributes and image identifications.
6. The electromagnetic scan-based industrial robot path optimization system according to claim 1, wherein the constraint generation module comprises a model receiving unit, an obstacle recognition unit, an obstacle boundary construction unit, a gradient analysis unit, a posture limit generation unit, and a constraint construction unit; The model receiving unit receives a space environment model, wherein the space environment model comprises image boundary labels, electromagnetic signal amplitude values, electromagnetic gradient values and corresponding space coordinates of all three-dimensional grid units; the obstacle recognition unit screens three-dimensional grid units with image boundary labels in the space environment model, and three-dimensional grid units with electromagnetic signal amplitudes exceeding a set threshold, and defines a central coordinate set of the screened three-dimensional grid units as an obstacle point set; The obstacle boundary construction unit executes a three-dimensional geometric envelope algorithm on the obstacle point set to generate a minimum convex boundary body containing all obstacle points; The gradient analysis unit calculates the gradient change rate of electromagnetic signals between any adjacent grid units in the space environment model, marks a corresponding grid area with the gradient change rate of the electromagnetic signals exceeding a preset threshold value as an electromagnetic interference area, and records the space coordinate range of the interference area in the three-dimensional space; the gesture restriction generating unit combines the space position of the electromagnetic interference area and the coordinate axis direction marked in the space environment model, calculates the allowable gesture orientation range of the end effector in the corresponding space area, and defines the space gesture which does not meet the corresponding gesture orientation as an infeasible area; The constraint construction unit maps the space coordinate range of the minimum convex boundary body, the space coordinate range of the electromagnetic interference area and the infeasible area direction into a convex constraint form in a path search variable space, and constructs a dynamic convex constraint set comprising geometric obstacle constraint, electromagnetic strength constraint and attitude limitation constraint; Wherein, the geometrical obstacle constraint means that the path point cannot fall into the polyhedron defined by the convex hull; electromagnetic intensity constraints mean that the waypoints are not allowed to occur in areas where the gradient intensity exceeds a threshold; pose constraint means that the pose vector of a path point within a particular region must belong to a set of allowed directions.
7. The electromagnetic scanning-based industrial robot path optimization system according to claim 1, wherein the path evolution module comprises an initial path construction unit, a differential disturbance generation unit, a constraint feedback adjustment unit, a cross combination unit and a validity screening unit; the initial path construction unit generates a plurality of path initial solutions meeting convex constraint conditions in a path search variable space; For each initial path solution, the differential disturbance generating unit constructs a path differential vector based on a plurality of other path solutions, synthesizes the differential vector with a basic path, and generates a first variant path candidate solution; After the first variation path candidate solution is generated, the constraint feedback adjustment unit receives the update information of the convex constraint set, judges the trend of the path point deviating from the feasible region according to the update information, and dynamically corrects the differential disturbance direction to obtain a second variation path candidate solution; Selecting two or more different path individuals from the current population, calculating vector differences between path points corresponding to the path individuals, weighting the difference vectors, and adding the weighted difference vectors as disturbance items to another path individual to generate a new second variant path candidate solution; The cross combination unit randomly combines the current path initial solution and the second variation path candidate solution at the path point level to generate a cross path solution fusing the original structure and the disturbance structure; And the validity screening unit performs convex constraint condition inspection on each path point in the cross path solution, identifies a path point set meeting the constraint condition, eliminates infeasible path points and constructs a continuous structure to obtain a final variant path candidate solution.
8. The electromagnetic scan based industrial robot path optimization system of claim 7, wherein the variant path candidate solution comprises a plurality of consecutive path points, each path point having coordinate information in a path search space.
9. The electromagnetic scanning-based industrial robot path optimization system according to claim 1, wherein the path constraint module comprises a path candidate receiving unit, a constraint set loading unit, a path point analyzing unit, a path projection computing unit and a path reconstructing unit; The path candidate receiving unit receives a variant path candidate solution; the constraint set loading unit receives a dynamic convex constraint set of the current optimization round, wherein the convex constraint set comprises a plurality of types of environment constraints formed by obstacle boundaries, electromagnetic interference areas and gesture direction limitations; The path point analysis unit analyzes the path points in the variant path candidate solution one by one, maps each path point into a constraint space defined by a convex constraint set, and establishes a constraint corresponding relation between the path points and the constraint set; The path projection operation unit executes gradual constraint projection operation on each path point according to the constraint corresponding relation, and sequentially maps the path points to all convex constraint sets until the spatial distance between two successive updates of the path points is lower than a preset threshold; and the path reconstruction unit gathers and sorts the converged path point results, and reconnects the converged path point results according to the space sequence of the original path to generate a complete path structure in the range of the convex constraint set.
10. The electromagnetic scan based industrial robot path optimization system of claim 1, wherein the path output module comprises an adaptability evaluation unit, a path optimization unit; the adaptability evaluation unit evaluates each path in the path set output by the path constraint module, wherein the evaluation comprises a path length index, an electromagnetic interference accumulation index, an attitude deviation index, a smoothness index and a constraint proximity index; The path length index is the total movement distance of the path in the three-dimensional space; The electromagnetic interference accumulation index is the exposure degree of the path in the electromagnetic interference area; The gesture deviation index is an average deviation angle between the tail gesture and the target gesture of the path in the execution process; the smoothness index is the angle change rate of the path between adjacent sections; the constraint proximity index is the minimum distance distribution from each path point in the path to the constraint boundary; and the path optimization unit calculates the adaptive total score of each path by adopting a weighted comprehensive scoring mode according to the adaptive index, and determines the path with the highest score as the optimal path as the target path for the control execution of the industrial mechanical arm.

Description

Industrial mechanical arm path optimization system based on electromagnetic scanning Technical Field The invention relates to the technical field of industrial automation and intelligent control, in particular to an industrial mechanical arm path optimization system based on electromagnetic scanning. Background With the accelerated development of intelligent manufacturing and industrial automation, industrial mechanical arms play a central role in various scenes such as welding, carrying, spraying and the like. At present, intelligent planning of a mechanical arm movement path in a complex environment becomes one of key technologies for guaranteeing operation efficiency and safety, a common path planning method generally depends on a fixed geometric model or a simple obstacle detection mechanism, a static environment modeling means based on vision is adopted, and a traditional evolutionary algorithm or a heuristic search algorithm is combined to perform path optimization. The existing path planning technology still has obvious limitations in processing dynamic and unstructured electromagnetic environments. On the one hand, the existing system is often used for constructing a static environment model only based on image or laser radar data, and is difficult to sense physical factors such as electromagnetic interference, invisible obstacle or abnormal field intensity which influence the stable operation of the mechanical arm, so that the path planning is easy to be interfered in a real environment, and the problems of poor robustness, incomplete identification of a feasible region and the like exist. On the other hand, the traditional evolutionary algorithm such as the differential evolutionary algorithm lacks effective fusion of real-time environment constraint in the path optimization process, the conditions of path point drift, unsatisfied obstacle avoidance or attitude limitation and the like are easy to occur in the optimization process, and particularly, the constraint failure problem is outstanding in the multidimensional space path evolution. Therefore, how to provide an industrial robot path optimization system based on electromagnetic scanning is a need to be solved. Disclosure of Invention The invention aims to provide an industrial mechanical arm path optimization system based on electromagnetic scanning, which comprises an image recognition module, a scanning control module, an environment modeling module, a constraint generation module, a path constraint module and a path output module; The image recognition module acquires an image of a target area through an image acquisition device arranged on the industrial mechanical arm, and extracts image boundary information and space positioning data of the target area according to color characteristics; the scanning control module controls the industrial mechanical arm to move to a target area according to the space positioning data, and performs near-field electromagnetic sampling through an electromagnetic scanning device at the tail end of the industrial mechanical arm to acquire original electromagnetic data; the environment modeling module performs joint modeling on the image boundary information and the original electromagnetic data to construct a space environment model for path planning; The constraint generation module extracts obstacle information and physical boundary conditions related to path planning based on a space environment model, and forms a dynamically updated convex constraint set; the path evolution module generates a plurality of path initial solutions meeting convex constraint conditions in a path search variable space, and performs differential evolution operation on the path initial solutions to obtain a plurality of variant path candidate solutions; the path constraint module calls a POCS algorithm to demap variant path candidates into a convex constraint set to obtain a path set meeting environmental constraints; And the path output module selects an optimal path according to the adaptability indexes of the paths of the path set which are determined to meet the environmental constraint, and converts the optimal path into a control instruction to execute the path action of the mechanical arm. Further, the image recognition module comprises an image acquisition unit, an image conversion unit, a parameter setting unit, a mask generating unit, a boundary extraction unit and a coordinate analysis unit; The image acquisition unit acquires color image frame data of a target area through a camera device arranged at the tail end of the industrial mechanical arm; the image conversion unit converts color image frame data into HSV color space images; The parameter setting unit stores a parameter set for color segmentation; the mask generating unit extracts pixel areas which accord with set color characteristics from the HSV color space image according to the parameter set and generates a mask image containing a target area; when a pixe