CN104965485A - Path planning method of surface array salient point on-demand jet printing control system - Google Patents

Abstract

A path planning method of a surface array salient point on-demand jet printing control system belongs to the technical field of on-demand jet printing, and aims to solve the problem that at present motion path planning efficiency is low in printing of salient points. The method includes: 1. reading in data to be printed, the number of cycles k=1; 2. judging whether k is smaller than x, if yes, outputting a current calculated shortest path and distance, and if no, departing from a starting point to visit, and turning to 3; 3. moving from a current point to the next allowable selection point, using an ant colony algorithm to calculate a path, and turning to 4; 4. judging whether the next allowable selection point exists for the current point, if yes, turning to 3, and if no, turning to 5; 5. judging whether the current ant has walked through all points, if yes, calculating the length of the path, and turning to 6, and if no, setting the length of the path to be infinite, and turning to 6; 6. judging whether the acquired length of the path is the shortest among existing paths, if yes, updating the shortest path and length, and turning to 7, and if no, turning to 7; and 7. updating pheromones of each side on the shortest path, plus 1 to k, and turning to 2. The path planning method of the surface array salient point on-demand jet-printing control system is used for surface array salient point printing.

Description

The paths planning method of face array salient point need based jet print control system

Technical field

The invention belongs to need based jet printing technique field.

Background technology

The cardinal principle of need based jet printing technique is after forcing melted material to be sprayed by nozzle with the form of droplet with vibration, by accurately controlling the motion of workbench, work in coordination with the course of injection of droplet and control the parameter such as droplet deposition position, droplet flying distance, liquid droplets deposition the forming surface interconnected salient point of array microelectronics Packaging and jet printing circuit etc.

For the technical requirement of microelectronics Packaging, the different interconnected stud bump making device of domestic and international software engineering researchers invent, the Central China University of Science and Technology utilizes ink-jet printer control technology to establish a set of piezoelectric type to spray 3 D-printing system; Jilin University utilizes motion control card and linear electric motors to establish the kinematic system of Piezoelectric Driving spray site glue ^[3], there is higher kinematic accuracy; Shanghai Communications University adopts kinematic error compensation technology, establishes the BGA laser reballing system based on Visual C++ platform ^[4].But still carry out path planning for the distributed intelligence of Area array packages solder bump at present, thus realize the kinetic control system of printing flexibly.

Print this application for face array salient point, the trajectory path planning problem of salient point in print procedure can be summed up as travelling salesman's class problem (TSP), belongs to Combinatorial Optimization category.For such problem, have traditional heuritic approaches such as the method for exhaustion, dynamic programming algorithm, genetic algorithm at present, but these Algorithm for Solving efficiency are too low.

Summary of the invention

The object of the invention is to there is the too low problem of efficiency to solve the trajectory path planning of current salient point in print procedure, the invention provides the paths planning method of a kind of array salient point need based jet print control system.

The paths planning method of of the present invention array salient point need based jet print control system,

Described method comprises the steps:

Step one: the node coordinate data of reading in n processing to be printed, its set is N, cycle index k=1;

Step 2: judge whether cycle index k is less than x, if so, then export shortest path and the distance thereof of current calculating, this method terminates, and if not, then conducts interviews from the starting point of face array salient point, proceeds to step 3;

Step 3: move to next step point allowing to select from current point, utilize ant group algorithm to carry out calculating path, proceed to step 4; The described point selected that allows is gather the node do not selected in N;

Step 4: judge whether also there is next step point allowing to select in current point, if the step 3 of proceeding to, proceed to step 5 if not;

Step 5: judge that whether current ant has passed by all points, if so, then calculates the length of this paths, proceeds to step 6; If not, then the length in this path is set to infinity, proceeds to step 6;

Step 6: whether the length in the path that determining step five obtains is shorter than the shortest existing path, if so, then upgrades shortest path and length, proceeds to step 7; If not, then step 7 is proceeded to;

Step 7: the pheromones upgrading each limit on shortest path, cycle index k adds 1, proceeds to step 2.

In described step 3, the method utilizing ant group algorithm to carry out calculating path is:

The order of ant according to 1≤s≤m is calculated respectively, when ant is at some i, a j access is judged, determines path:

$j=\left\{\begin{array}{ccc}\arg & \underset{l\∈J_k}{\max }\left[{\left({\mathrm{\τ}}_{\mathrm{il}}\right)}^{\mathrm{\α}}{\left({\mathrm{\η}}_{\mathrm{il}}\right)}^{\mathrm{\β}}\right]& q\≤q_0\\ & P_{\mathrm{ij}}^k& q>q_0\end{array}\right.$

Wherein, J _k={ N-t _k, represent the point that next step permission of ant is selected, t _kbe called taboo list, be used for recording ant k all points of passing by; α is called information heuristic greedy method, and β is expected heuristic value, and q is equally distributed random number on [0,1] interval;

The limit value q of node transition rule ₀span be 0≤q ₀≤ 1, $P_{\mathrm{ij}}^k=\left\{\begin{array}{cc}\frac{({\left[{\mathrm{\τ}}_{\mathrm{ij}}\right]}^{\mathrm{\α}}\·{\left[{\mathrm{\η}}_{\mathrm{ij}}\right]}^{\mathrm{\β}})}{{\mathrm{\Σ}}_{l\∈J_k}{\left[{\mathrm{\τ}}_{\mathrm{il}}\right]}^{\mathrm{\α}}\·{\left[{\mathrm{\η}}_{\mathrm{il}}\right]}^{\mathrm{\β}}}& j\∈J_k\\ 0& j\∉J_k\end{array}\right.;$

Total number m=n, the η of ant in ant group _ijrepresent that the quantity of information of path (i, j) is τ by the expected degree of an i to some j _ij.

In described step 7, the method upgrading the pheromones on each limit on shortest path is:

Pheromones vestige on shortest path W is strengthened, to the pheromones τ on other paths _ijk () is volatilized:

${\mathrm{\τ}}_{\mathrm{ij}}\left(k\right)=\left\{\begin{array}{cc}(1-{\mathrm{\ρ}}_{k-1}){\mathrm{\τ}}_{\mathrm{ij}}(k-1)+\frac{{\mathrm{\ρ}}_{k-1}}{L\left(W\right)}& (i,j)\∈W\\ (1-{\mathrm{\ρ}}_{k-1}){\mathrm{\τ}}_{\mathrm{ij}}(k-1)& (i,j)\∉W\end{array}\right.;$

Wherein, volatilization factor ρ _kmeet following relation:

$\left\{\begin{array}{c}{\mathrm{\ρ}}_k\≤1-\frac{\ln k}{\ln (k+1)}\\ \underset{k=1}{\overset{\∞}{\mathrm{\Σ}}}{\mathrm{\ρ}}_k=\∞\end{array}\right.---\left(5\right).$

Beneficial effect of the present invention is, the present invention is based on ant group algorithm, and prints for face array salient point and improve, and adds feedback mechanism, makes result of calculation more level off to absolute optimal solution.

In method provided by the invention, the ant of global optimum is only had just to be allowed to release pheromone, its objective is in the field of the best path found out till the search of guarantee ant mainly concentrates on previous cycle, ant effectively can be avoided to converge to same path, and decrease the number of times without efficient search.Adopt the mechanism of pheromones volatilization, object is the searching probability in order to increase the limit that those are not accessed to, and contributes to the expansion of region of search, prevents system to be absorbed in local minimum too early, makes system be provided with degenerative function.And have employed pheromones and strengthen process and can realize the centralized action that cannot be realized by single ant, the pheromones achieved on optimal path strengthens, and makes system be provided with the function of positive feedback.Whole planning system can be made like this to have positive and negative feedback function, improve the precision of system, also the search of system is optimized.And at the end of algorithm is final, ant group remembers the bee-line of optimal path up to the present and correspondence.

Paths planning method provided by the invention can before kinetic control system setting in motion, path planning is carried out according to user coordinates data, move again after obtaining a path that distance is the shortest, the time that motion process needs can be reduced, improve the efficiency that salient point prints.

Accompanying drawing explanation

Fig. 1 is the principle schematic of the face array salient point print system described in embodiment.

Fig. 2 is the principle schematic of the kinetic control system described in embodiment.

Fig. 3 is the schematic flow sheet of the paths planning method of the face array salient point need based jet print control system described in embodiment.

Fig. 4 is the schematic diagram adopting method of the present invention to carry out moving line before path planning in embodiment.

Fig. 5 is the schematic diagram adopting method of the present invention to carry out moving line after path planning in embodiment.

Embodiment

Present embodiment also provides a face array salient point print system, and described array salient point print system comprises main control computer, drive waveform system, droplet generator, kinetic control system and three-dimensional movement platform;

After face array salient point need based jet print routine in main control computer starts, first all data in user coordinates file will be read, carry out path planning process and form new coordinate file, then read the data point coordinate in new coordinate file successively and drive three-dimensional movement platform motion by kinetic control system, judge whether to reach terminal, end program after reaching home, otherwise just move according to the difference of coordinate.By drive waveform system, droplet generator is controlled simultaneously, and carry out two system coordinations controls, print to realize face array salient point.

The principle schematic of kinetic control system as shown in Figure 2.

In this kinetic control system, because the interconnected salient point of face array integrated circuit layout on circuit boards exists in the form of an array, when the given salient point coordinate time needing to print, motion control card can along regulation coordinate sequential movements, before printing, main control computer needs to process motion path, finds a route being convenient to move.Utilize Grid Method to obtain the grating map of three-dimensional movement platform, using data given for user as node coordinate, need to find one to access each point to be printed at least one times, and can ensure that total distance is shorter, planning comparatively reasonably path.

Composition graphs 3 illustrates present embodiment, the paths planning method of the face array salient point need based jet print control system described in present embodiment,

010, open the coordinate data file that user provides, read in the node coordinate data of all processing to be printed, be set to set N, the number at number of users strong point is n.

020, for face array salient point print platform, due to diaxon self-movement separately, therefore the distance definition between 2 is the absolute value sum of the absolute value of the difference of the horizontal ordinate of 2 and the difference of ordinate.Suppose that the ant number that t is positioned at an i is a _ii (), the quantity of information in t path (i, j) is τ _ijt (), i point is to the distance d of j point _ijrepresent.For native system, in setting ant group, total number m of ant equals the number at number of users strong point, i.e. m=n.If the set of the point that ant s walks is L (s), time initial, L (s) is empty set, 1≤s≤m.K represents whole algorithm cycle index, η _ijrepresent by the expected degree of an i to some j, be referred to as heuristic function, determined by formula (1):

${\mathrm{\η}}_{\mathrm{ij}}\left(t\right)=\frac{1}{d_{\mathrm{ij}}}---\left(1\right)$

030, judge whether cycle index k is less than x, the x value in present embodiment is 500.X can determine according to required computational accuracy, computing time of limiting.X value is larger, and the degree of accuracy of final calculation result is higher, and required computing time is longer.Described degree of accuracy refers to the difference of result of calculation and actual shortest path;

040, judge whether whether meet cycle index k is less than 500, if met, then stop calculating, export shortest path and the distance thereof of current calculating; Otherwise, then make ant s conduct interviews from the starting point of face array salient point, proceed to step 050.

050, next step point allowing to select is moved to from current point, ant group algorithm is utilized to carry out calculating path: the order of ant according to 1≤s≤m to be calculated respectively, when ant is at some i, L (s) is judged, if L (s) ≠ N, so just carry out conducting interviews judgement to a j according to formula (2):

$j=\left\{\begin{array}{ccc}\arg & \underset{l\∈J_k}{\max }\left[{\left({\mathrm{\τ}}_{\mathrm{il}}\right)}^{\mathrm{\α}}{\left({\mathrm{\η}}_{\mathrm{il}}\right)}^{\mathrm{\β}}\right]& q\≤q_0\\ & P_{\mathrm{ij}}^k& q>q_0\end{array}\right.---\left(2\right)$

Wherein, J _k={ N-t _k, represent the point that next step permission of ant is selected, t _kbe called taboo list, be used for recording ant k all points of passing by.α, β are given systematic parameters, the power of representative information element and 2 distances select the influence degree of next coordinate to ant respectively, and wherein α is called information heuristic greedy method, and its value is larger, the path that ant more easily selects other ants to pass by, value is 1 in the present embodiment; β is expected heuristic value, and reflect ant by the impact of distance factor on the judgement of any under it, value is 2 in the present embodiment; Q is equally distributed random number on [0,1] interval.Q ₀(0≤q ₀≤ 1) be the limit value of node transition rule of setting, its size determine utilization before relative importance between result and Probing new way footpath, get q in the present embodiment ₀=0.9.As q≤q ₀time, this ant is then according to result of calculation selecting paths before, otherwise this ant just carries out the exploration to new route according to formula (3):

$P_{\mathrm{ij}}^k=\left\{\begin{array}{cc}\frac{({\left[{\mathrm{\τ}}_{\mathrm{ij}}\right]}^{\mathrm{\α}}\·{\left[{\mathrm{\η}}_{\mathrm{ij}}\right]}^{\mathrm{\β}})}{{\mathrm{\Σ}}_{l\∈J_k}{\left[{\mathrm{\τ}}_{\mathrm{il}}\right]}^{\mathrm{\α}}\·{\left[{\mathrm{\η}}_{\mathrm{il}}\right]}^{\mathrm{\β}}}& j\∈J_k\\ 0& j\∉J_k\end{array}\right.$

(3); Proceed to step 060;

060, judge whether also there is next step point allowing to select, i.e. J in current point _kwhether be empty set, if J _k{ j} repeats step 050, if J to ≠ Φ, then L (s)=L (s) ∪ _k=Φ, then after access, L (s)=L (s) ∪ { i ₀, proceed to step 070.

070, judge ant s whether passed by a little, namely whether L (s) equals N, if L (s)=N, then according to the order calculating path total length of L (s) mid point, be designated as f (L (s)), proceed to step 080; If L (s) ≠ N, be then set to infinity by path, proceed to step 080;

080, judge that in 070, whether path total length is shorter than existing shortest path, namely the path of m ant is compared, the ant that note wherein walks shortest path is r, the shortest path this time obtained before search is W (f (W) is initialized as infinity), if f (L (t)) <f (W), then make W=L (t); Proceed to step 090;

090, utilize formula (4) to strengthen the pheromones vestige on W path, the pheromones on other paths volatilized:

${\mathrm{\&tau;}}_{\mathrm{ij}}\left(k\right)=\left\{\begin{array}{cc}(1-{\mathrm{\&rho;}}_{k-1}){\mathrm{\&tau;}}_{\mathrm{ij}}(k-1)+\frac{{\mathrm{\&rho;}}_{k-1}}{L\left(W\right)}& (i,j)\&Element;W\\ (1-{\mathrm{\&rho;}}_{k-1}){\mathrm{\&tau;}}_{\mathrm{ij}}(k-1)& (i,j)\&NotElement;W\end{array}\right.---\left(4\right)$

Wherein, volatilization factor ρ _kfor fixing cycle index k>=1, meet following relation:

$\left\{\begin{array}{c}{\mathrm{\&rho;}}_k\&le;1-\frac{\ln k}{\ln (k+1)}\\ \underset{k=1}{\overset{\&infin;}{\mathrm{\&Sigma;}}}{\mathrm{\&rho;}}_k=\&infin;\end{array}\right.---\left(5\right)$

Formula (5) illustrates if a limit is not optimal path, and so after k volatilization, its pheromones is gradually reduced to disappearance.According to formula above, new τ can be obtained _ijk (), makes k=k+1, repeat 040.

Pheromones τ _ijk () is an array, represent and comprise multiple quantity of information τ _ij.

Embodiment of the method:

The paths planning method provided in the present invention is provided, utilizes Matlab coding, generate M file, then utilize Visual C++ that M file is compiled as dll file, calling in other programming softwares can be realized.

Utilize the kinetic control system control program provided in the present invention, realize the exploitation of the kinetic control system of face array salient point print platform, development platform selects LabVIEW, in kinetic control system, carry out calling path planning procedure.

First control system carries out the planning of motion path after running, and then generates shortest path and preserves.After reading the coordinate data on shortest path successively, motion platform moves.

Run this control system, what first carry out is the planning of motion path, Fig. 4 and Fig. 5 illustrates and a kind ofly to seesaw the contrast of route carrying out path planning, Fig. 4 represents motion path when not carrying out path planning according to salient point distributed data, and can calculate total distance of moving in motion process from figure is 73 coordinate unit; Fig. 5 represents the motion path after carrying out path planning, and total distance that can calculate motion is 39 coordinate unit.In this instance, distance when not planned by the travel after path planning decreases 46%, effectively improves efficiency.

Paths planning method provided by the invention not only can utilize Matlab to realize, and other programming software (as Visual C++ etc.) also can be utilized to realize.

The implementation that Platform for Motion Control provided by the invention is not limited only on hardware " computing machine+motion control card+servomotor ", alternate manner such as the mode such as " computing machine+PLC+ servomotor ", " single-chip microcomputer+servomotor " also can realize motion control function.On software, be not limited only to take LabVIEW as development platform, other development platform such as visual c++ etc. also can realize the function of kinetic control system exploitation.

Paths planning method provided by the invention not only can be used in face array salient point and print, in the Quick-forming that can be used in other equally or panel path plan optimization problem.

Paths planning method provided by the invention includes but not limited to that workbench is the situation of two-dimension moving platform, forms need based jet print system that relative two dimensional moves all within the scope of the invention between all jetting system and workbench.

Claims (3)

1. the paths planning method of an array salient point need based jet print control system, it is characterized in that, described method comprises the steps:

Step one: the node coordinate data of reading in n processing to be printed, its set is N, cycle index k=1;

Step 2: judge whether cycle index k is less than x, if so, then export shortest path and the distance thereof of current calculating, this method terminates, and if not, then conducts interviews from the starting point of face array salient point, proceeds to step 3;

Step 3: move to next step point allowing to select from current point, utilize ant group algorithm to carry out calculating path, proceed to step 4; The described point selected that allows is gather the node do not selected in N;

Step 4: judge whether also there is next step point allowing to select in current point, if the step 3 of proceeding to, proceed to step 5 if not;

Step 5: judge that whether current ant has passed by all points, if so, then calculates the length of this paths, proceeds to step 6; If not, then the length in this path is set to infinity, proceeds to step 6;

Step 6: whether the length in the path that determining step five obtains is shorter than the shortest existing path, if so, then upgrades shortest path and length, proceeds to step 7; If not, then step 7 is proceeded to;

Step 7: the pheromones upgrading each limit on shortest path, cycle index k adds 1, proceeds to step 2.

2. the paths planning method of according to claim 1 array salient point need based jet print control system, is characterized in that, in described step 3, the method utilizing ant group algorithm to carry out calculating path is:

The order of ant according to 1≤s≤m is calculated respectively, when ant is at some i, a j access is judged, determines path:

$j=\left\{\begin{array}{cc}\arg \underset{l\&Element;J_k}{\max }\left[{\left({\mathrm{\&tau;}}_{\mathrm{il}}\right)}^{\mathrm{\&alpha;}}{\left({\mathrm{\&eta;}}_{\mathrm{il}}\right)}^{\mathrm{\&beta;}}\right]& q\&le;q_0\\ P_{\mathrm{ij}}^k& q>q_0\end{array}\right.$

Wherein, J _k={ N-t _k, represent the point that next step permission of ant is selected, t _kbe called taboo list, be used for recording ant k all points of passing by; α is called information heuristic greedy method, and β is expected heuristic value, and q is equally distributed random number on [0,1] interval;

The limit value q of node transition rule ₀span be 0≤q ₀≤ 1, $P_{\mathrm{ij}}^k=\left\{\begin{array}{cc}\frac{({\left[{\mathrm{\&tau;}}_{\mathrm{ij}}\right]}^{\mathrm{\&alpha;}}\&CenterDot;{\left[{\mathrm{\&eta;}}_{\mathrm{ij}}\right]}^{\mathrm{\&beta;}})}{{\mathrm{\&Sigma;}}_{l\&Element;J_k}{\left[{\mathrm{\&tau;}}_{\mathrm{il}}\right]}^{\mathrm{\&alpha;}}\&CenterDot;{\left[{\mathrm{\&eta;}}_{\mathrm{il}}\right]}^{\mathrm{\&beta;}}}& j\&Element;J_k\\ 0& j\&NotElement;J_k\end{array}\right.;$

Total number m=n, the η of ant in ant group _ijrepresent that the quantity of information of path (i, j) is τ by the expected degree of an i to some j _ij.

3. the paths planning method of according to claim 2 array salient point need based jet print control system, is characterized in that, in described step 7, the method upgrading the pheromones on each limit on shortest path is:

Pheromones vestige on shortest path W is strengthened, to the pheromones τ on other paths _ijk () is volatilized:

${\mathrm{\&tau;}}_{\mathrm{ij}}\left(k\right)=\left\{\begin{array}{cc}(1-{\mathrm{\&rho;}}_{k-1}){\mathrm{\&tau;}}_{\mathrm{ij}}(k-1)+\frac{{\mathrm{\&rho;}}_{k-1}}{L\left(W\right)}& (i,j)\&Element;W\\ (1-{\mathrm{\&rho;}}_{k-1}){\mathrm{\&tau;}}_{\mathrm{ij}}(k-1)& (i,j)\&NotElement;W\end{array}\right.;$

Wherein, volatilization factor ρ _kmeet following relation:

$\left\{\begin{array}{c}{\mathrm{\&rho;}}_k\&le;1-\frac{\ln k}{\ln (k+1)}\\ \underset{k=1}{\overset{\&infin;}{\mathrm{\&Sigma;}}}{\mathrm{\&rho;}}_k=\&infin;\end{array}\right.---\left(5\right)$

。