CN-115239788-B - Multi-target optimization method for movement track of powder extraction robot

CN115239788BCN 115239788 BCN115239788 BCN 115239788BCN-115239788-B

Abstract

The invention discloses a multi-target optimizing method for a moving track of a powder extraction robot, which solves the problem that the moving track is not accurate enough when the powder extraction robot extracts, and the technical scheme is that the method comprises the following steps that S1, information of a powder target area is acquired, an extracted feasible area is constructed, and environment initialization is carried out; and S3, solving the multi-target optimization model in the step S2 by adopting a decomposition multi-target evolutionary algorithm based on improvement, and optimizing and extracting the moving track of the robot so as to improve the extraction efficiency and accuracy.

Inventors

LIU WEIWEI
WANG SHENGNAN
ZHAO CHENTONG
FANG WEI
Qi shuo
HUANG XIAOYUE
XU WEI

Assignees

沈阳工业大学

Dates

Publication Date: 20260508
Application Date: 20220408

Claims (7)

1. The multi-target optimizing method for the moving track of the powder extraction robot is characterized by comprising the following steps of: s1, acquiring powder object area information, constructing an extracted feasible area, and initializing the environment; S2, inputting the information after the initialization in the step S1 into a multi-objective optimization model; S3, solving the multi-target optimization model in the step S2 by adopting an improved decomposition multi-target evolutionary algorithm, and optimizing and extracting the moving track of the robot; in the step S2, the information after the initialization is input into the multi-objective optimization model, which specifically includes the following steps: First, according to the target powder area A and purity requirement Calculating the area of the extraction feasible region, planning an initial environment and track distribution nodes again, and establishing a multi-objective optimized mathematical model of the moving track of the powder extraction robot according to the area, purity and track nodes of the feasible region: Wherein: To optimize the actual extraction purity and target purity according to the moving track The difference between the two, Representing the sum of all road segments moved; representing the sum of included angles of two adjacent road sections after the optimization of the moving track; The actual extraction purity and target purity requirement after the optimization according to the moving track Optimal objective function of difference The method comprises the following steps: Optimal objective function of sum of all road segments of the movement The method comprises the following steps: The sum of the included angles of two adjacent sections of the moving track is optimal to an objective function The method comprises the following steps: Wherein: representing the number of updated first movement trajectory optimization nodes, Is the first The number of nodes in the network is, The diameter size of the suction nozzle of the robot, For the number of trips to and from the robot zone, Mobile track node Coordinates of (c); Therein is as follows Track length between; Is in the track The angle between two adjacent road segments at a point.
2. The multi-target optimizing method for the moving track of the powder extraction robot according to claim 1, wherein in the step S1, target area information is acquired, and the method for constructing the extracted feasible area comprises the steps of acquiring an image of a target powder area, performing image processing and analysis on the image, extracting the target powder area from a background area, performing binarization processing, obtaining a target powder area A, calculating the area of the extracted feasible area according to the target powder area A, and constructing the extracted feasible area according to the calculated area of the extracted feasible area.
3. The method for optimizing a movement trajectory of a powder extraction robot according to claim 1, wherein the environmental initialization in step S1 includes the steps of first establishing a global coordinate system 0-XY, initializing a range of X, Y axis direction constraints in a target powder region to be extracted and an extraction feasible region in the coordinate system, respectively, thereafter extracting N node positions of the trajectory, forming an initial node of a first walking trajectory in the target powder region to be extracted, and finally, updating the initial node of the first walking trajectory formed in the target powder region to be extracted in the feasible region.
4. The method for optimizing a movement trajectory of a powder extraction robot according to claim 1, wherein in step S3, the improved decomposition-based multi-objective evolutionary algorithm comprises the steps of: step 301, initializing the first extracted position of the mobile track node, then updating the node, and optimizing the node sequence; step S302, decomposing the target space and setting any first space The weight vector of the sub-problem is Wherein Is the first The weight that the individual objective function takes up, The weight vector for each sub-problem finds T nearest vectors, i.e. neighborhoods, according to the neighborhood size set in step S301 Improving chebyshev function and calculating node In the first place Synthesizing objective functions of the sub-problems, and then carrying out normalization operation; step S303, modeling each decomposed weight vector in step S302 by using a probability model, and randomly sampling to obtain a new solution ; Step S304 for each index Calculate the sampled in step S303 If the composite objective function value of (2) Then Quilt is covered with Instead of the above-mentioned, ; = As the vector of the best reference point, = Is the worst reference point vector, wherein, Is the target number; Step S305, terminating the iteration when the current iteration number is satisfied Maximum number of iterations = Stopping the calculation, otherwise Returning to step S303; Step S306, outputting target value Corresponding optimal solution N is N sequences and reference points, and the number of extraction round trips is set 。
5. The method of claim 4, wherein in step S301, the node update is specifically that in the region of the powder to be extracted, N initial node sequences are sampled to obtain more sufficient, and the nodes which finally form N updated first walking tracks are expressed as Wherein Is a sequence of discrete track points; initial iteration number Setting the maximum iteration number Randomly initializing population pop= = (number of sub-problems) of population scale N ) I.e. N sequences and reference points, a set of weight vectors is set Neighborhood size of sub-problem.
6. The method of claim 4, wherein in step S302, the modified chebyshev function is: s.t. where Ω is the decision space, function Is the first And an objective function.
7. The method according to claim 4, wherein in step S306, the first extracted movement track formed according to the optimal solution is used as the diameter of the nozzle in the feasible region The method is based on the following steps: , And Is the range of x-axis direction constraints.

Description

Multi-target optimization method for movement track of powder extraction robot Technical Field The invention belongs to the technical field of material separation devices, and particularly relates to a multi-objective optimization method for a moving track of a powder extraction robot. Background The powder extraction is an extraction technology aiming at powder intermediate products, is widely applied to the fields of petroleum, chemical industry, medicine, biochemistry and the like, and has higher difficulty in accurately extracting target objects because the boundaries of different compound layers are staggered and staggered. In laboratory micro-component analysis, manual extraction is usually adopted, but the extraction speed is low and the raw material utilization rate is low. At present, a robot is used for replacing manual operation, and although extraction efficiency can be improved, in the existing process of extracting powder by the robot, the moving track of the robot is set inaccurately, so that when the robot extracts the powder, the phenomenon that part of powder distribution areas are not extracted or part of areas are repeatedly extracted is easy to occur. Disclosure of Invention Aiming at the defects of the prior art, the invention aims to provide a multi-target optimization method for the moving track of a powder extraction robot, which solves the problem of unreasonable design of the robot on the extraction track of the powder, so that the robot can search and optimize more accurately, thereby improving the extraction efficiency and accuracy. The invention aims to realize the multi-objective optimization method of the moving track of the powder extraction robot, which comprises the following steps: s1, acquiring powder object area information, constructing an extracted feasible area, and initializing the environment; S2, inputting the information after the initialization in the step S1 into a multi-objective optimization model; and S3, solving the multi-target optimization model in the step S2 by adopting an improved decomposition multi-target evolutionary algorithm, and optimizing and extracting the moving track of the robot. Further, in the step S1, target area information is acquired, and the construction of the extracted feasible area comprises the steps of acquiring an image of a target powder area, performing image processing and analysis on the image, extracting the target powder area from a background area, performing binarization processing, obtaining the area A of the target powder area, calculating the area of the extracted feasible area according to the area A of the target powder area, and constructing the extracted feasible area according to the calculated area of the extracted feasible area. Further, in the step S1, the environment initialization comprises the steps of firstly establishing a global coordinate system 0-XY, initializing the range of constraint of x and y axes in a target powder area to be extracted and an extracted feasible area in the coordinate system respectively, then extracting N node positions of a track, forming an initial node of a first walking track in the target powder area to be extracted, and finally updating the initial node of the first walking track formed in the target powder area to be extracted in the feasible area. Further, in the step S2, the information after the initialization is input into the multi-objective optimization model, which specifically includes the following steps: Firstly, calculating the area of an extraction feasible region according to the area A of the target region and the purity requirement eta, then planning an initial environment and a distribution sub-track node N again, and establishing a multi-target optimized mathematical model of the moving track of the powder extraction robot according to the area, the purity and the track node of the feasible region: minF(x)=(f1(x),f2(x),f3(x)) Wherein f 1 (x) is the difference between the actual extraction purity and the target purity requirement eta after the optimization of the moving track, f 2 (x) is the sum of all the moving road sections, and f 3 (x) is the sum of the included angles of two adjacent road sections after the optimization of the moving track. Further, the optimal objective function f 1 (x) of the difference between the actual extraction purity and the target purity requirement η after the optimization according to the moving track is: the optimal objective function f 2 (x) of the sum of all the road segments of the movement is: The optimal objective function f 3 (x) of the sum of the included angles of two adjacent sections of the moving track is as follows: Wherein n represents the number of updated first movement track optimizing nodes, i is the ith node, ζ is the diameter size of a robot suction nozzle, μ is the round trip number of a robot region, (x i,yi),(xi+1,yi+1) is the coordinates of a movement track node P i,Pi+1, |P i,Pi+1 | is the track leng