TWM601358U - System of workpiece inspection in path planning - Google Patents

Abstract

This model provides a path planning system for workpiece inspection, which plans the best inspection path for workpiece inspection by measuring machine. A coordinate relationship matrix is established based on the movement mode of the ray inspection measuring machine between the two detection points and the complex movement mode of the detection points. The path optimization module encodes each detection point, and then arranges the detection points in a random order multiple times to become a complex detection path, and sorts the detection paths according to their total distance value. The mutation detection path is generated according to the evolution of the whale calculation unit, which includes the total distance value of the mutation. The total distance value and the total distance value of the mutation are compared and the smallest one is selected to be the best detection path. In this way, a fast and safe optimal path for workpiece detection is generated.

Description

Path planning system for workpiece inspection

This model relates to a path planning system, especially to a path planning system for workpiece detection.

In the process of precision manufacturing, when the workpiece is processed, it needs to be measured by a measuring machine to ensure that the precision of the processed workpiece is within a reasonable range. When the number of inspection points to be inspected increases, users cannot smoothly plan the measurement path based on experience. Therefore, if the path planning problem can be effectively solved, production efficiency can be significantly improved.

The current path planning uses the following methods: (1) Minimize the trajectory to reduce the measurement time in the Coordinate Measuring Machine (CMM), and use the ant colony algorithm to search. (2) The measurement sequence is planned by the nearest neighbor method and the refinement method, and the optimized sequence is used to complete the entire measurement process. The above-mentioned path planning method still has many problems, such as too fast convergence speed, long time required to solve, easy to fall into the best solution in the region, and huge amount of calculation and need to use a lot of memory to complete. It can be seen that there is currently no optimal solution in the market that has slower convergence, shorter solution time, and is not easy to fall into the region. The path planning system of China, so related businesses are seeking its solutions.

Therefore, the purpose of this new model is to provide a path planning system for workpiece inspection, which first detects the movement between the inspection points and generates a coordinate relationship matrix, encodes and randomly sorts the inspection points into a complex inspection path, and calculates the path of each inspection path. The total distance value, and each detection path is sorted according to the total distance value, and then the group is divided and the leader in each group is selected. Then use the whale calculus unit to generate the mutation detection path and the mutation total distance, and select the smallest total distance value from each leader and mutation detection path as the best detection path to solve the problem of the best solution that is easy to fall into the region in the conventional path planning technology problem.

According to one of the implementations of the structural aspect of the present invention, a path planning system for workpiece inspection is provided to plan the best inspection path for the measuring machine to detect the workpiece. The path planning system for workpiece inspection includes a coordinate relationship matrix establishment module and a path Optimization module. The coordinate relationship matrix establishment module drives the movement mode of the ray inspection measuring machine between the two detection points to generate the coordinate relationship matrix. Path optimization module signal connection coordinate relationship matrix building module, path optimization module includes initialization coding module, fitness value sorting module, dynamic grouping module, mutation module, selection module and iteration number judgment module group. The initial coding module is used to code each detection point, and then drive the detection points to be arranged in random order multiple times to generate multiple detection paths. The fitness value sorting module, the signal connection initialization coding module, is based on the coordinate relationship matrix and each detection path to generate the total distance value of each detection path, and according to the size of the total distance value Sort detection paths. The dynamic grouping module, the signal connection fitness sorting module, groups the detection paths into multiple groups according to the total distance value, and the difference between the total distance values of the multiple detection paths in any group is less than or equal to the preset difference. The mutation module, the signal is connected to the dynamic grouping module, the leader of the group generates a mutation detection path through the whale calculation unit, and the mutation detection path includes the total distance value of the mutation. Select the module, connect the signal to the mutation module, compare the total distance value of the leader in each group, the total distance value of the mutation, and the best detection path after the previous iteration, and select the smallest one to be the best detection path. The iteration number judgment module, the signal connects the selection module and the mutation module, the iteration number judgment module determines whether the execution times of the mutation module is equal to the preset number; if not, the mutation module and the selection module are executed again; if yes , The path optimization module terminates execution.

In this way, the path planning system of the new type of workpiece detection escapes the regional optimal solution through the dynamic grouping module and the whale calculation unit, so as to solve the problem that the conventional path planning system converges too fast and easily falls into the regional optimal solution.

Other examples of the foregoing embodiment are as follows: In the foregoing coordinate relationship matrix establishment module, if there is a workpiece between two detection points, the movement mode is non-linear movement. If there is no workpiece between the two detection points, the movement mode is linear movement.

Other examples of the foregoing embodiment are as follows: the foregoing dynamic grouping module includes a similarity calculation sub-module and a grouping sub-module. The similarity calculation sub-module calculates the suitability threshold and distance threshold of the detected path. The grouping sub-module, the signal connection similarity calculation sub-module calculates the difference between each detection path and the leader The fitness difference and distance difference are divided into groups according to the fitness threshold and the distance threshold.

Other examples of the foregoing embodiment are as follows: the foregoing whale calculation unit includes an enclosing sub-module, a hunting sub-module, and a random search sub-module. Surround the sub-module to update the best detection path after the previous iteration and find the smallest total distance value of each leader. Hunting sub-module, the signal is connected to surround the sub-module to find the smallest total distance value of each leader. Random search sub-modules, signal connection to hunting sub-modules, arranging detection points in random order to generate mutation detection paths.

Other examples of the aforementioned implementation are as follows: the aforementioned workpiece inspection path planning system further includes a re-encoding module, a signal connection iteration number judging module, and the re-encoding module converts the encoding of the best inspection path into a measuring machine for use The code of the measuring machine.

200: Path planning system for workpiece inspection

210: Coordinate relationship matrix creation module

212: Path Optimization Module

214: Recode module

220: Initialize the coding module

230: fitness value sorting module

240: Dynamic grouping module

242: Similarity calculation submodule

244: Group sub-module

250: mutation module

252: Whale Calculus Unit

254: Surround submodule

256: Hunting submodule

258: Random Search Submodule

260: Select Module

270: Iteration number judgment module

100: Path planning method for workpiece inspection

152: Whale Algorithm

S110: Steps for establishing coordinate relationship matrix

S112: Path optimization step

S114: Recoding step

S120: Initial coding step

S130: fitness value sorting steps

S140: Dynamic grouping steps

S142: step of calculating similarity

S144: Grouping steps

S150: mutation step

S154: Bracketing step

S156: Hunting Step

S158: Random search step

S160: Selection steps

S170: Steps to determine the number of iterations

Figure 1 is a block diagram of the path planning system for workpiece inspection in the first embodiment of the present invention; Figure 2 is a block diagram of the path optimization module of the path planning system for workpiece inspection in Figure 1; Figure 3 is a block diagram of the dynamic grouping module of the path optimization module of Figure 2; Figure 4 is a block diagram of the whale calculation unit of the path planning system for workpiece detection in the first embodiment of the present invention Schematic diagram Figure 5 is a schematic flow diagram of the path planning method for workpiece inspection according to the second embodiment of the present invention; Figure 6 is a schematic flow diagram of the path optimization steps of the path planning method for workpiece inspection in Figure 5; Figure 7 is a schematic flow diagram of the dynamic grouping step of the path optimization step in Figure 6; and Figure 8 is a schematic flow diagram of the whale algorithm of the path planning method for workpiece detection in the second embodiment of the present invention.

Please also refer to Figure 1. Figure 1 is a block diagram of the workpiece inspection path planning system 200 according to the first embodiment of the present invention. As shown in the figure, the workpiece inspection path planning system 200 planning measurement machine The best inspection path for inspecting workpieces. The path planning system 200 for workpiece inspection includes a coordinate relationship matrix establishment module 210, a path optimization module 212, and a recoding module 214.

The coordinate relationship matrix establishment module 210 is used to drive the movement of the ray inspection and measurement machine between the two detection points to generate the coordinate relationship matrix. If there is a workpiece between the two detection points, the movement mode is non-linear movement; if there is no workpiece between the two detection points, the movement mode is linear movement. In detail, whether the ray detection probe can move in a straight line between two detection points, it is divided into "0" and "1" according to the way of movement between points, where "0" represents between two points It can be moved in a straight line. "1" means that the two points cannot be moved in a straight line. The Z-axis needs to be moved safely to avoid collisions. And the movement mode between the two detection points is established as a coordinate relationship matrix.

Please refer to FIG. 2. FIG. 2 is a block diagram of the path optimization module 212 of the path planning system 200 for workpiece detection in FIG. 1. FIG. Path optimization module 212 signal connection coordinate relationship matrix establishment module 210. Path optimization module 212 includes initialization coding module 220, fitness value sorting module 230, dynamic grouping module 240, mutation module 250, selection The module 260 and the iteration number judgment module 270. In other words, the path optimization module 212 is used to list the detection points in a plurality of arrangements and find the best detection path from them.

The initial coding module 220 is used to code each detection point, and then drive the detection points to be arranged in a random order multiple times to generate multiple detection paths. In other words, the initialization encoding module 220 encodes the coordinates of each detection point and generates a plurality of detection paths arranged in random order.

Please refer to Table 1. The fitness value sorting module 230 is connected to the initialization coding module 220. The fitness value sorting module 230 generates the total distance value of the detection path according to the coordinate relationship matrix and each detection path, and sorts according to the total distance value. Detect paths, and preset the group of each detection path to 0. The total distance value is Fitness, and each detection path is I _n . Sort I _n according to its fitness from minimum to maximum, I ₁ is the smallest total distance value, I _NP is the largest total distance value, and Group is the group.

Please refer to FIG. 3 in conjunction. FIG. 3 is a block diagram of the dynamic grouping module 240 of the path optimization module 212 in FIG. 2. As shown in the figure, the dynamic grouping module 240 is connected to the fitness sorting module 230 to group the detection paths into multiple groups according to the total distance value. The difference between the total distance values of the multiple detection paths in any group is less than or equal to The preset difference. The dynamic grouping module 240 includes a similarity calculation sub-module 242 and a grouping sub-module 244.

In detail, the similarity calculation sub-module 242 calculates the suitability threshold and the distance threshold of the detected path. The calculation method of the similarity calculation submodule 242 conforms to the following formula:

其中DIS^g為第g組的距離差，n為檢測路徑的總數，D為檢測點的總數，g為組別，

為第g組的領袖，

為第i個檢測路徑，NC為當前組別的檢測路徑總數，φ^g為第g組的距離閾值。FIT^g為第g組的適合度差，Fit(τ^g)為第g組的領袖之適應度值，Fit(Xⁱ)為第i個檢測路徑的適應度值，ψ^g為第g組的適合度閾值。

Where DIS ^g is the distance difference of the g-th group, n is the total number of detection paths, D is the total number of detection points, and g is the group,

Is the leader of group g,

Is the i-th detection path, NC is the total number of detection paths in the current group, and φ ^g is the distance threshold of the g-th group. FIT ^g is the fitness difference of the g- ^th group, Fit(τ ^g ) is the fitness value of the leader of the g- ^th group, Fit(X ⁱ ) is the fitness value of the i-th detection path, and ψ ^g is the fitness value of the g-th group The fitness threshold.

The grouping sub-module 244 signal connection similarity calculation sub-module 242 calculates the fitness difference and distance difference between each detection path and the leader, and divides them into groups according to the fitness threshold and the distance threshold. The grouping submodule 244 conforms to the following formula:

為第g組的領袖，

為第i個檢測路徑，Fit(τ^g)為第g組的領袖之適應度值，Fit(Xⁱ)為第i個檢測路徑的適應度值。當Disⁱ<φ^g和Fitⁱ<ψ^g時，表示第i個檢測路徑與第g組領袖是相似的，第i個檢測路徑會被歸入第g組，所有檢測路徑皆有組別後，分群子模組244執行終止執行。 Fit ⁱ =| Fit (τ ^g )- Fit ( X ⁱ )| (6). Where Dis ⁱ is the distance difference of the i-th detection path, Fit ⁱ is the fitness difference of the i-th detection path, D is the total number of detection points,

Is the leader of group g,

Is the i-th detection path, Fit(τ ^g ) is the fitness value of the leader of the g- ^th group, and Fit(X ⁱ ) is the fitness value of the i-th detection path. When Dis ⁱ <φ ^g and Fit ⁱ <ψ ^g , it means that the i-th detection path is similar to the g-th group leader, and the i-th detection path will be classified into the g-th group, and all detection paths have groups. , The grouping sub-module 244 executes termination execution.

Please refer to FIG. 4, which is a block diagram of the whale calculation unit 252 of the path planning system 200 for workpiece detection according to the first embodiment of the present invention. The mutation module 250 is signaled to the dynamic grouping module 240. The mutation module 250 generates a mutation detection path through the whale arithmetic unit 252 for the leaders of these groups. The mutation detection path includes the total distance value of the mutation. The whale calculation unit 252 includes an enclosure sub-module 254, a hunting sub-module 256, and a random search sub-module 258.

The surrounding sub-module 254 updates the best detection path after the previous iteration and finds the smallest total distance value of each leader. The surrounding submodule 254 conforms to the following formula:

其中t為當前的迭代次數，X(t)為當前的最佳檢測路徑，X^*(t)為檢測路徑的位置空間，X_τg為第g組之領袖，A、C為係數，其計算方法如第(9)式及第(10)式，a為進行迭代時從2線性下降到0的向量，r表示0到1之間的隨機數。換句話說，包圍子模組254載入前次迭代後的最佳檢測路徑及各領袖之總距離值之最小者。

Where t is the current iteration number, X(t) is the current best detection path, X ^* (t) is the location space of the detection path, X _τg is the leader of the g-th group, A and C are coefficients, and its calculation method Such as (9) and (10), a is a vector that linearly decreases from 2 to 0 when iterating, and r represents a random number between 0 and 1. In other words, the surrounding submodule 254 loads the best detection path after the previous iteration and the smallest total distance value of each leader.

其中p為0到1之間的隨機數，b為用來定義螺旋形狀的常數，1為1到-1之間的隨機數。當p<0.5時，以收縮環繞方式找出各領袖之總距離值最小之檢測路徑；當p>0.5時，以螺旋運動方式找出各領袖中總距離值最小之檢測路徑。 The hunting sub-module 256 signals the surrounding sub-module 254 to find the smallest total distance value of each leader. The hunting submodule 256 conforms to the following formula:

Where p is a random number between 0 and 1, b is a constant used to define the spiral shape, and 1 is a random number between 1 and -1. When p<0.5, the detection path with the smallest total distance value of each leader is found by shrinking and wrapping; when p>0.5, the detection path with the smallest total distance value of each leader is found by spiral motion.

The random search sub-module 258 is connected to the hunting sub-module 256 with a signal to arrange the detection points in a random order to generate a mutation detection path. The random search submodule 258 conforms to the following formula:

其中X_rand為隨機產生的檢測路徑之向量，⊕表示乘法， D向量的產生方式導入Lévy飛行策略，Lévy飛行策略從提供的概率分佈中提供隨機游動，使鯨魚演算單元252能跳脫區域最佳解。當A

1時，隨機產生一個突變檢測路徑，突變檢測路徑包含突變總距離值。

Where X _rand is the vector of the detection path generated randomly, ⊕ represents multiplication, and the method of generating D vector is imported into the Lévy flight strategy. The Lévy flight strategy provides random walks from the probability distribution provided, so that the whale calculation unit 252 can escape the area most Good solution. When A

At 1:00, a mutation detection path is randomly generated, and the mutation detection path contains the total distance value of the mutation.

其中

為前次迭代後的最佳檢測路徑，

為隨機搜索子 模組258所產生的突變檢測路徑。選擇模組260判斷

的總距離值是否比

更好，若是，則

取代

成為最佳檢測路徑；若否，則前次迭代後的最佳檢測路徑保留至下一代。 The signal of the selection module 260 is connected to the mutation module 250, and the total distance value of the leader in each group, the total distance value of the mutation, and the best detection path after the previous iteration are compared, and the smallest remaining one is selected to become the best detection path. The selection module 260 conforms to the following formula:

among them

Is the best detection path after the previous iteration,

The mutation detection path generated by the sub-module 258 is randomly searched. Selection module 260 judgment

Is the total distance value greater than

Better, if yes, then

replace

Become the best detection path; if not, the best detection path after the previous iteration is retained to the next generation.

The iteration number determination module 270 signals the connection selection module 260 and the mutation module 250. The iteration number determination module 270 is used to determine whether the execution times of the mutation module 250 are equal to the preset number; if not, the mutation module 250 and the selection module 250 The module 260 is executed again; if so, the path optimization module 212 terminates execution. In other words, when the iteration reaches the preset number of times, the obtained optimal inspection path is the optimal inspection path requested by the path planning system 200 for workpiece inspection.

The re-encoding module 214 signals the connection iteration number judging module 270. The re-encoding module 214 converts the encoding of the best detection path into The code of the measuring machine used by the measuring machine. In detail, the measuring machine can be a computer numerical control (CNC) machine, and the measuring machine code used by the measuring machine can be G/M code.

In this way, the path planning system 200 for workpiece detection of the present invention escapes the regional optimal solution through the dynamic grouping module 240 and the whale calculation unit 252, so as to solve the problem that the conventional path planning system converges too fast and easily falls into the regional optimal solution. problem.

Please refer to Figures 1 to 6 together. Figure 5 shows the flow diagram of the workpiece inspection path planning method 100 according to the second embodiment of the present invention; Figure 6 shows the workpiece inspection of Figure 5 A schematic flowchart of the path optimization step S112 of the path planning method 100. As shown in the figure, the path planning method 100 for workpiece inspection is used to plan the best inspection path for the measuring machine to inspect the workpiece, and the inspection path includes a plurality of inspection points. The path planning method 100 for workpiece detection includes a coordinate relationship matrix establishment step S110, a path optimization step S112, and a recoding step S114.

The coordinate relationship matrix establishment step S110 is to drive the movement mode of the radiation inspection and measuring machine between the two detection points, and establish the coordinate relationship matrix according to the movement mode of the detection points. If there is a workpiece between the two detection points, the movement mode is non-linear movement; if there is no workpiece between the two detection points, the movement mode is linear movement. The coordinate relationship matrix creation step S110 is performed through the coordinate relationship matrix creation module 210.

The path optimization step S112 includes an initialization coding step S120, an fitness ranking step S130, a dynamic grouping step S140, a mutation step S150, a selection step S160, and an iteration number determination step S170.

The initial coding step S120 is to code each detection point, and then arrange the detection points in a random order multiple times to form a complex detection path. The initialization coding step S120 is performed by the initialization coding module 220.

The fitness sorting step S130 is to calculate the total distance value of each detection path according to the coordinate relationship matrix and each detection path. The fitness value sorting step S130 is performed by the fitness value sorting module 230.

Please refer to FIG. 7, which is a schematic flowchart of the dynamic grouping step S140 of the path optimization step S112 in FIG. 6. The dynamic grouping step S140 is to group the detection paths into multiple groups according to the total distance value. The difference between the total distance values of the multiple detection paths in any group is less than or equal to the preset difference, and the total distance value in each group is The smallest is regarded as the leader. The dynamic grouping step S140 includes a similarity calculation step S142 and a grouping step S144. The dynamic grouping step S140 is executed through the dynamic grouping module 240.

The similarity calculation step S142 is to calculate the suitability threshold and the distance threshold of the detected path. The similarity calculation step S142 is performed through the similarity calculation sub-module 242.

The grouping step S144 is to calculate the fitness difference and distance difference between each detection path and the leader, and divide into groups according to the fitness threshold and the distance threshold. The grouping step S144 is executed by the grouping sub-module 244.

Please refer to FIG. 8. FIG. 8 is a flowchart of the whale algorithm 152 of the path planning method 100 for workpiece inspection according to the second embodiment of the present invention. The mutation step S150 is to evolve the leader of the group according to the whale algorithm 152 to generate a mutation detection path. The mutation detection path includes the mutation detection path. Change the total distance value. The whale algorithm 152 includes a surrounding step S154, a hunting step S156, and a random search step S158.

The encircling step S154 is to update the best detection path after the previous iteration and find the smallest value of the total distance of each leader in a shrinking and encircling manner. The enclosing step S154 is performed through the enclosing sub-module 254.

It is judged whether the value of the random random number p is less than 0.5, if it is, the encircling step S154 is executed again; if not, the hunting step S156 is executed. The hunting step S156 is to find the smallest total distance value of each leader in a spiral motion. The hunting step S156 is executed by the hunting sub-module 256.

The random search step S158 is to arrange the detection points in a random order to generate a mutation detection path. The random search step S158 is performed by the random search sub-module 258.

The selection step S160 is to compare the total distance value of the leader in each group, the total distance value of the sudden change, and the best detection path after the previous iteration, and select the one that remains the smallest to become the best detection path. The selection step S160 is performed through the selection module 260.

The iteration number determination step S170 is to determine whether the execution number of the mutation step S150 is equal to the preset number; if not, the mutation step S150 and the selection step S160 are executed again; if so, the path optimization step S112 is ended. The iteration number determination step S170 is executed by the iteration number determination module 270.

The re-encoding step S114 is to convert the code of the best detection path into the code of the measuring machine used by the measuring machine. The re-encoding step S114 is performed by the re-encoding module 214.

In this way, the new path planning method 100 for workpiece detection escapes the regional optimal solution through the dynamic grouping step S140 and the whale algorithm 152, so as to solve the problem that the conventional path planning method has too fast convergence speed and easily falls into the regional optimal solution. .

Although the present invention has been disclosed in the above embodiments, it is not intended to limit the present invention. Anyone who is familiar with this technique can make various changes and modifications without departing from the spirit and scope of the present invention. Therefore, the protection of the present invention The scope shall be subject to those defined in the attached patent scope.

200: Path planning system for workpiece inspection

210: Coordinate relationship matrix creation module

212: Path Optimization Module

214: Recode module

Claims (5)

A path planning system for workpiece inspection, which is used to plan the best inspection path for a workpiece by a measuring machine. The path planning system for workpiece inspection includes: a standard relationship matrix establishment module that drives a ray to detect the measurement The machine moves between two detection points to generate a standard relationship matrix; and a path optimization module, the signal is connected to the coordinate relationship matrix establishment module, the path optimization module includes: an initialization code The module is used to encode the detection points, and then drive the detection points to be arranged in a random order multiple times to generate a complex detection path; a fitness value sorting module, the signal is connected to the initialization coding module, based on the coordinate relationship matrix And each detection path generates a total distance value of each detection path, and sorts the detection paths according to the size of the total distance values; a dynamic grouping module, the signal is connected to the fitness value sorting module, according to the total distance value The distance value groups the detection paths into a complex group, and the difference between the plurality of the detection paths in any one group and the total distance value is less than or equal to a preset difference; a mutation module, the signal is connected to the dynamic grouping Module, the plural leaders of the groups generate a mutation detection path through a whale calculation unit, the mutation detection path includes a total mutation distance value; a selection module, the signal is connected to the mutation module, and the group is The The total distance value of the leader, the total distance value of the sudden change, and the best detection path after the previous iteration are compared and the smallest one is selected to be the best detection path; and an iteration number judgment module, the signal is connected to the selection Module and the mutation module, the iteration number determination module determines whether an execution number of the mutation module is equal to a preset number; if not, the mutation module and the selection module are re-executed; if so, then The path optimization module terminates execution. The path planning system for workpiece detection according to claim 1, wherein in the coordinate relationship matrix establishment module, if the workpiece exists between the two detection points, the movement mode is non-linear movement; and if the two detection points If the workpiece does not exist between, the movement method is linear movement. The path planning system for workpiece inspection according to claim 1, wherein the dynamic grouping module includes: a similarity calculation sub-module for calculating a suitability threshold and a distance threshold of the detection paths; and a grouping sub-module A module, a signal connected to the similarity calculation sub-module, calculates a fitness difference and a distance difference between each detection path and the leader, and divides them into the groups according to the fitness threshold and the distance threshold. The path planning system for workpiece detection according to claim 1, wherein the whale calculation unit includes: an enveloping sub-module that updates the best detection path after the previous iteration and finds the total distance value of each leader The smallest A hunting sub-module with a signal connected to the surrounding sub-module to find the smallest value of the total distance of each leader; and a random search sub-module with a signal connected to the hunting sub-module, arranged in random order These detection points generate the mutation detection path. The path planning system for workpiece inspection according to claim 1, further comprising: a re-encoding module, the signal is connected to the iteration number judgment module, and the re-encoding module converts the encoding of the best inspection path into the quantity The measuring machine uses a measuring machine code.

Images