CN103471592A - Multi-unmanned aerial vehicle route planning method based on bee colony collaborative foraging algorithm - Google Patents

Abstract

The invention belongs to the field of unmanned aerial vehicle technology, and more specifically relates to a multi-unmanned aerial vehicle route planning method which can be used for multi-unmanned aerial vehicle target tracking, route optimization, cooperative management, cooperative flight, and task distribution. The multi-unmanned aerial vehicle route planning method comprises following steps: detailed information of route planning task is initialized, and coordinate discrete transformation is performed; parameters of the bee colony collaborative foraging algorithm is initialized; route cost of each unmanned aerial vehicle is calculated; current route cost is calculated according to a current position by each unmanned aerial vehicle; guide aerial vehicles are selected by followed aerial vehicles, wherein one followed aerial vehicle is recruited by each guide aerial vehicle; the current route is abandoned and a new route is searched; parameters of optimized route are saved, and an optimal value is calculated; inspection is performed so as to determine whether an upper limit of iteration number is reached. The multi-unmanned aerial vehicle route planning method is capable of resolving the initialization sensitive problem of a traditional intelligent optimization route planning method, and can be used for improvement of accuracy of the algorithm so as to obtain an optimal solution or a second best solution sufficiently close to the optimal solution, and improve stability of route planning and search efficiency.

Description

A kind of multiple no-manned plane path planning method based on the collaborative algorithm of looking for food of bee colony

Technical field

The invention belongs to the unmanned plane technical field, be specifically related to a kind of multiple no-manned plane path planning method that can be used for multiple no-manned plane target following, flight path optimization, coordinated management, collaborative flight, task distribution.

Background technology

Under future Information Battle Environment complicated and changeable, single frame unmanned plane will be difficult to finish the work, and the multiple UAVs that must fly by Collaborative Control in a lot of situations just can complete; Each unmanned plane requires 1 to 3 people's crew to distribute, and consults and coordinate many mankind soldiers.Except the cost of human operator who, this method runs into indeterminable challenge, how to reach collaborative.Under the restriction of current science and technology, want unmanned machine to arrive pilot's powerful information processing capability and intelligence or difficult like that, if by the clustering phenomenon of natural imitation circle biology, the unmanned plane quantitatively had comparative advantage utilizes swarm intelligence just can reach even to surmount the people that has who quantitatively accounts for inferior position drives machine.The Evolution of analysis of biological system and behavior rule, combine some principle of biotic population intelligence and behavior with multiple no-manned plane Collaborative Control theory, have wide future in engineering applications.Though the research of the collaborative flight of unmanned aerial vehicle group at present and trajectory planning aspect has obtained certain achievement in research at home and abroad, also there is no unified theory and effective method.

Trajectory planning is according to known various environmental informations, comprises terrain information and enemy's situation information, considers all sidedly the restriction of unmanned plane each side ground, avoids the full spectrum of threats environment, cooks up rational flight track, makes UAV complete safely preplanned mission.The domestic and international research of the offline mode about colony's unmanned plane at present mainly concentrates on Task Allocation Problem modeling aspect, the achievement in research of present stage mainly comprises many traveling salesman problems (Mutiple Traveling Salesman Problem, MTSP) model, Vehicle Routing Problems (Vehicle Routing roblem, VRP) model, MILP (Mixed Integer Linear Programming) (Mixed Integer Linear Programming, MILP) model etc.

Path Planning is broadly divided into following two large classes:

1, deterministic calculation.Deterministic calculation have based on minimal principle and A ^*algorithm; Heuristic search A* algorithm advantage is that convergence is strong quick etc. with computing, and shortcoming is that it can only generate a flight path, is not suitable for the mission requirements of many reference tracks of those needs.

2, Stochastic search optimization algorithm, include simulated annealing, genetic algorithm and ant group algorithm etc.The thought that this class algorithm adopts is the material change procedure of simulating nature circle, also has the process of biological activity and evolution, has many advantage and disadvantages, therefore also is widely used on the trajectory planning problem.Because the binding character of search volume can not limit genetic algorithm, do not need the continuity of majorized function and the conditions such as existence of derivative yet, also there is concurrency, relatively be applicable to the trajectory planning problem that comprises much complex constraint and fuzzy message; And ant group algorithm adapts to the polytrope of threatening environment more owing to having dynamic perfromance.

Although possess, the intelligent algorithms such as genetic algorithm, particle cluster algorithm and ant group algorithm of random characteristics are of overall importance and locality is all fine, can also generate a plurality of solutions, can overcome the shortcomings and limitations of first kind method to a certain extent, but calculated amount is larger, speed of convergence is slower, therefore often is difficult to meet the actual needs of engineering.

Voronoi figure method be take method of geometry as foundation, and overall characteristics of planning is better, and is not suitable for the planning of local tracks, and shortcoming is exactly the polytrope that can only do two dimensional surface planning and be not suitable with battlefield surroundings.

In order to overcome the shortcoming of said method, constantly there are some new methods that naturally inspire to propose.

Such as having described, the people such as Madhavan utilize geometrical rule to produce the paths planning method of unmanned aerial vehicle group cooperation, referring to Madhavan S., Antonios T., Brian W.Co-operative path planning of multiple UAVs using Dubins paths with clothoid arcs.Control Engineering Practice18 (2010) 1084 – 1092.The people such as Marinakis have proposed the extensive paths planning method based on honeybee mating optimized algorithm, referring to Marinakis Y, Marinaki M, Dounias G.Honey bees mating optimization algorithm for large scale vehicle routing problems[J] .Natural Computing, 2010,9 (1): 5-27.The people such as Duan Haibin have proposed the unmanned aerial vehicle flight path plan optimization algorithm based on chaos bee colony optimized algorithm, referring to " Xu, Chunfang, Duan, Haibin; Liu, Fang.Chaotic artificial bee colony approach to Uninhabited Combat Air Vehicle (UCAV) path planning.Aerospace Science and Technology, 14 (8), p535-541,2010 ".Hu Zhonghua etc. utilize ant colony algorithm to carry out single unmanned aerial vehicle flight path planning, referring to Hu Zhonghua, and Zhao Min. the unmanned aerial vehicle flight path planning based on the artificial bee colony algorithm, research sensor and micro-system, the 3rd phase: 1-5 in 2010.

At present, although some above-mentioned classic methods are arranged and comprise that novel intelligent optimization algorithm that bee colony optimizes is for the unmanned aerial vehicle flight path planning problem, and obtained certain achievement in research, but traditional unmanned plane cluster task distributes modeling method from mathematical modeling angle research problems, does not realize the advantage of colony of social insect; The methods such as above-mentioned swarm intelligence and evolutionary computation do not have again the real control that goes out to send to realize the unmanned plane cluster from the natural essence of colony of social insect, and all not for the multiple no-manned plane trajectory planning.All do not have to solve the key issues such as the aloft Collaborative Control of unmanned plane cluster, trajectory planning, collision prevention from the angle of simulation bee group biology behavior, only the angle from optimizing, obtain abstract solution, limited to the practical problems effect.

The bee colony intelligent use, in unmanned plane cluster offline mode research and comparison novelty, is had to very important researching value and meaning

Summary of the invention

The object of the invention is to propose a kind of multiple no-manned plane path planning method based on the collaborative algorithm of looking for food of bee colony that improves the collaborative efficiency of executing the task of multiple no-manned plane and security, reliability.

The object of the present invention is achieved like this:

Multiple no-manned plane path planning method based on the collaborative algorithm of looking for food of bee colony, comprise the steps:

(1), according to environmental model, the details of initialization trajectory planning task, carry out the coordinate discrete transform;

(2) the collaborative algorithm parameter of looking for food of initialization bee colony, comprise threshold value T search time, and iterations is K, Bas=0 search time of initialization algorithm, and initial iterations K=1 population total N, lead the machine number N ₁with follow the machine number N ₂, N=N ₁+ N ₂;

(3), according to the unmanned aerial vehicle flight path parameter, calculate the cost f in the flight path path of every unmanned plane ^j(j=1,2,3..., n),

x (j) is j bar air route, J _i(x (j)) is i item routeing target penalty function in the routeing target, ω _ifor the weight coefficient of each penalty, j is the unmanned aerial vehicle flight path ordinal number, and n is the flight path sum, the total item that k is needs assessment in evaluation function;

(4) every unmanned plane, according to current position calculation current path cost, if the current path cost is lower than original route cost, upgrades the position of unmanned plane: for i unmanned plane, be created in the integer j of [1, D], [1, NE] integer k, j parameter of i unmanned plane used

change, the unmanned aerial vehicle flight path that calculates the new cost value after undated parameter and select a low-cost is as new flight path, unmanned plane position vector eigenwert, j ∈ 1,2 ..., Q}, rand (0,1) is (0,1) upper equally distributed random number, x _max, x _minrespectively x _ihigher limit and lower limit;

(5) according to the cost value of every unmanned plane, the machine of following selects to lead object, and each leads machine to recruit one and follows machine, leading the machine surrounding space to continue the search new route, repeating step (4), if this new route is lower than original route cost, that is converted into the machine that leads at random, upgrade the position of every unmanned plane, continue to survey, search time, Bas set to 0 again, if this new route is higher than original route cost, the maintenance search, Bas adds 1;

(6) if Bas search time is greater than definite critical value, unmanned plane is abandoned current path, and the search new route, re-execute (2);

(7) preserve the parameter of optimal path and calculate optimal value;

(8) check whether reach the iterations upper limit, reach and finish search, otherwise repeating step (4) is to step (7).The details of initialization trajectory planning task, carry out the coordinate discrete transform, comprises i navigation spots N _icoordinate be:

$\left\{\begin{array}{c}{x}_i=x_{i-1}+r\cos \mathrm{\θ}\cos \mathrm{\α}-r\sin \mathrm{\θ}\sin \mathrm{\α}\\ y_i=y_{i-1}+r\cos \mathrm{\θ}\sin \mathrm{\α}+r\sin \mathrm{\θ}\cos \mathrm{\α}\end{array}\right.,r\≥\frac{d(s,g)}{N},r\≥{4l}_{\min },$

S, g are respectively starting point and the terminal in initialization air route, and r is single air route section step-length, the air line distance that d (s, g) is the air route Origin And Destination, and N is the navigation spots number that air route allows at most, 4l _minfor the minimum flight air route segment length of unmanned plane constraint, (x _i, y _i) be the unmanned plane position coordinates, the angle that θ is unmanned plane route and transverse axis, α is the anglec of rotation.

Beneficial effect of the present invention is:

The present invention makes a plurality of unmanned planes search for path separately simultaneously, and the path of finding thus has global property, thereby has overcome the initialization tender subject of traditional intelligence optimization path planning method; Because employing leads search, follows search, follow and transform the Information Sharing Strategy renewal unmanned aerial vehicle flight path position led, the present invention can improve the precision of algorithm, in order to obtain optimum solution, or the suboptimal solution very approaching with optimum solution, promote the stability of trajectory planning, improved search efficiency; The present invention is owing to not adopting Deterministic rules, but probability of use rule guidance search, without grasping other prioris, has splendid practicality and fabulous robustness.

The accompanying drawing explanation

Fig. 1 is the FB(flow block) of performing step of the present invention;

Fig. 2 is the collaborative trajectory planning model of unmanned plane;

Fig. 3 is unmanned plane concerted attack direction schematic diagram;

Fig. 4 is node generation method schematic diagram in air route;

Fig. 5 selects the navigation spots of avoiding threatening as initialization navigation spots schematic diagram by tangent method;

Fig. 6 is with the simulation experiment result figure that does not adopt the collaborative path planning method image of bee colony to obtain;

Fig. 7 is with the simulation experiment result figure that adopts the collaborative path planning method of bee colony to obtain.

Embodiment

Below in conjunction with accompanying drawing, the present invention is described further:

The technical solution used in the present invention is that the collaborative foraging behavior of bee colony is introduced in the multiple no-manned plane trajectory planning to reach better trajectory planning performance, and for how man-machine trajectory planning Design of Problems lead search, follow search strategy, propose to follow search and be converted into the solution space update method that leads search, solve the high problem of computation complexity, obtain new multiple no-manned plane path planning method.Its specific implementation process is as follows:

(1), according to environmental model, the details of initialization trajectory planning task, carry out the coordinate discrete transform relevant to task;

(2) parameter of the collaborative algorithm of looking for food of initialization bee colony, comprise population total N, leads the machine number N ₁with follow the machine number N ₂, meet following condition: N=N ₁+ N ₂;

(3), according to the unmanned aerial vehicle flight path parameter, calculate the cost f in the flight path path of every unmanned plane ^j(j=1,2,3..., n), j is the unmanned aerial vehicle flight path ordinal number, and n is the flight path sum, and cost is less, and path is better;

(4) every its current path cost of position calculation that unmanned plane is current according to them, if the current path cost is lower than original cost, upgrade their position;

(5), according to the cost value of every unmanned plane, the machine of following selects to lead object.Each leads machine to recruit one and follows machine, is leading the machine surrounding space to continue the new solution of search, repeating step 4, and calculate its cost.If this new departure is better than original scheme, the machine of following is converted into the machine that leads so, upgrades the position of every unmanned plane, and continues to survey, and search time, Bas set to 0 again, otherwise, the maintenance search, Bas adds 1;

(6) if Bas search time is greater than definite critical value, unmanned plane will be abandoned current scheme, and again search for new route, return to (2);

(7) preserve the parameter of preferred embodiments and calculate optimal value;

(8) check whether reach the iterations upper limit, reach termination routine, otherwise repeating step 4-7.

With reference to Fig. 1, specific implementation step of the present invention is as follows:

Step 1, initial track modeling, coordinate transform.

(1) the heuristic initial method of improved end points

In order to guarantee the fully random of path profile, adopt end points to inspire initialized method to generate each navigation spots in air route, as shown in Figure 4.

$r\≥\frac{d(s,g)}{N}$

r≥4l _min

Wherein s, g are respectively starting point and the terminal in initialization air route.

I navigation spots N _icoordinate formula generate:

$\left\{\begin{array}{c}{x}_i=x_{i-1}+r\cos \mathrm{\θ}\cos \mathrm{\α}-r\sin \mathrm{\θ}\sin \mathrm{\α}\\ y_i=y_{i-1}+r\cos \mathrm{\θ}\sin \mathrm{\α}+r\sin \mathrm{\θ}\cos \mathrm{\α}\end{array}\right.$

In formula, r is single air route section step-length, and this parameter is suitably chosen meeting under above-mentioned two formula prerequisites, the air line distance that in formula, d (s, g) is the air route Origin And Destination, and N is the navigation spots numbers that air route allows at most, 4l _minminimum flight air route segment length for the unmanned plane constraint.It is according to need to the sample fourth class branch of air route section of the calculating of air route threat value that selection is greater than the minimum air route segment length of four times.θ is N in Fig. 4 _i-1the angle of g and transverse axis, α is an anglec of rotation.How to rotate and become the key that can the initialization population effectively avoid threat, choosing of α value will be an emphasis that improves the heuristic initialization of end points air route population study.

This method, by tangent method, selects the navigation spots of avoiding as far as possible threatening as air route initialization navigation spots.At first the navigation spots (on the straight line that this navigation spots is last navigation spots and impact point) that judgement is expanded, as the N in Fig. 4 _iwhether, in threatening area, if the expansion navigation spots is not threatening district expanding node not to be taked to operation, directly join in the individual chained list in air route, otherwise be rotated and make it break away from as far as possible threat according to certain rule.As shown in Figure 5, p' is last navigation spots, and p is the two kind situations of expansion navigation spots in the threat district.

Its cathetus l is vertical with straight line p ' p, α is the less angle between p ' p and p ' of mistake and circle o tangent line, the angle that β is p ' p and p ' o, θ be p ' o ' and cross p ' with circle o ' tangent line between angle, γ be p ' o ' ' with the angle between p ' and circle o ' ' tangent line.Can determine that thus the anglec of rotation is to break away from threat, make navigation spots p enter regional A or regional B.While for forward extent, only having single threat, rotation angle is α or alpha+beta, for the forward extent shown in Fig. 5, has the situation of a plurality of threats to try to achieve the anglec of rotation according to angular relationship.During rotation, rule adopts the random sense of rotation of selecting, and counterclockwise or clockwise direction, at first select the little anglec of rotation to carry out tentative rotation, after rotation, for different situations, does different disposal.

(1) if the new navigation spots produced after rotation not in threatening area and little anglec of rotation α be not more than 90 degree, select at random interval [α, 90] in, an angle is rotated again, if after rotation, navigation spots is not threatening district, this navigation spots is added to the individual chained list in air route, then again expand navigation spots, to the last the air line distance of navigation spots and impact point is less than 4l _min.Finally before impact point, add a navigation spots that meets the airfield runway vector.

(2) if the new navigation spots produced after rotation in former threatening area, is selected the large anglec of rotation (as Fig. 5 the first situation is selected α+2 β angles), if this angle is greater than 90 degree, change sense of rotation, adopting uses the same method does tentative rotation.If be less than 90 degree, in interval [α+2 β, 90], select an angle to be rotated, if do not break away from threat while rotating, the anglec of rotation is approached to α+2 β, make p ' p and former threat tangent.

(3) if can not find the interval of (1) (2) situation in rotary course, the random selection counterclockwise and making p ' p and threatening the tangent but minimum anglec of rotation in threatening not in clockwise direction, make p ' p and threat tangent but not in threat.In initialization procedure, the rotation of navigation spots being limited in 90 degree, guaranteeing that air route is forward direction, is the direction approximation towards impact point thereby make the navigation spots of air route individuality all the time.During rotation, sense of rotation is not limited, because all may find optimal air line from both direction

Step 2, the collaborative offline mode trajectory planning of bee colony, the parameter of initialization algorithm, comprise population total N, leads the machine number N ₁with follow the machine number N ₂, meet following condition: N=N ₁+ N ₂.Generally speaking, establish N ₁=N ₂, N=50 is set.Maximum restriction T search time, iterations is K.The initialization scope of activities, unmanned aerial vehicle flight path is expressed as Q dimensional vector x _i.Bas=0 search time of initialization algorithm, initial iterations K=1.

Step 3, according to the unmanned plane parameter, calculate the cost of every paths by the correlation parameter of formula (1), cost is less, path is better;

$f^j(w,x\left(j\right))=\underset{i=1}{\overset{k}{\mathrm{\Σ}}}{\mathrm{\ω}}_i{J}_i\left(x\left(j\right)\right)---\left(1\right)$

X in formula (j) is j bar air route, J _i(x (j)) is i item routeing target penalty function in the routeing target, ω _iweight coefficient for each penalty.The total item that K is needs assessment in evaluation function.

Step 4, in order to find new route scheme, unmanned plane calculates their current positions, if new flight path cost is lower than original cost, according to formula (2), upgrades their position;

This search strategy is described below: for i unmanned plane, at first immediately produce one [1, D] random integer j, [1, NE] a random integers k, then j the formula for parameter (2) of i unmanned plane changed, calculate the new cost value after undated parameter and select one than the unmanned aerial vehicle flight path of low-cost as new flight path.

$x_i^j=x_{\min }^j+\mathrm{rand}\left(\mathrm{0,1}\right)(x_{\max }^j-x_{\min }^j)---\left(2\right)$

unmanned plane position vector eigenwert, j ∈ in formula 1,2 ..., Q}, rand (0,1) is (0,1) upper equally distributed random number x _max, x _minrespectively x _ihigher limit and lower limit.

Step 5, according to the calculated value of unmanned aerial vehicle flight path, lead machine to adopt formula (3) to select the machine of following.Each leads machine to recruit one and follows machine, is leading the machine surrounding space to continue the new path node of search according to formula (4).Search for complete, repeating step 4, and calculate its cost.

$P_i={\mathrm{fit}}_i/\underset{i=1}{\overset{N}{\mathrm{\Σ}}}{\mathrm{fit}}_i---\left(3\right)$

Fit in formula (3) _ibe i unmanned aerial vehicle flight path cost, N is the unmanned plane sum.To probability, adopt the normalization mode to process herein, total probability and be 1.

J ∈ in formula 1,2 ..., Q}, k ∈ 1 ..., and i-1, i+1 ..., N}, k is random the generation, and

for a number random between [1,1].

If new route is lower than original route cost, follow so the machine flight path and be converted into and lead the machine flight path, upgrade the position of unmanned plane, and continue search, Bas sets to 0 again, otherwise, the maintenance search, Bas adds 1.

If step 6 Bas search time is greater than definite critical value, unmanned plane will be abandoned current path, and again search for new path node, reinitialize parameter and calculation cost.

Step 7, the parameter of preserving preferred embodiments and calculating optimal value.

Whether step 8, inspection reach the iterations upper limit, reach termination routine, otherwise repeating step 4-7.

Effect of the present invention can further illustrate by following emulation:

1. simulated conditions and emulation content:

The map size of procedure simulation is 780X620, has 10 and threatens point, and coordinate is respectively (304,400), (404,320), (440,500), (279,310), 560,220), (172,527), (194,220), (284,231), (447,242), (272,522); The threat radius is respectively: 45505510706555252025; Coordinate of ground point is (400,400), altogether three unmanned planes; The starting point coordinate is: (90,90), (40,600), (660,80); Attack is spaced apart 0, arrives simultaneously; Unmanned plane during flying velocity range 1.50～3m/s, 60 ° of hard-overs; Procedure Selection guiding honeybee quantity is 10, and following honeybee quantity is 100, on the iteration of simulations number of times, is limited to 20 times.

2. emulation experiment content

The emulation of multiple no-manned plane paths planning method of the present invention

A. multiple no-manned plane be take and attacked angle and be 120 °

It is that 120 ° of simulation results are as follows that angle is attacked in input.Before optimizing, the random path generated is as Fig. 6:

The same target of three airplane concerted attacks requires three airplanes to arrive the target overhead from three directions simultaneously.Three fly

Machine reference position coordinate is respectively (90,90), and (40,600), (660,80), coordinate of ground point is (400,400), it is 120 ° that three unmanned planes are attacked angle.

Utilize the method for trajectory planning, it is constraint that the 120 ° of angle directions of simultaneously take respectively arrive, and three unmanned planes are worked in coordination with to trajectory planning.Before optimization, flying distance is respectively: 484.446272m, 527.499636m, 495.164875m; Obtain the estimated time of arrival (ETA) t of team after collaborative _α=286s, three airplane flying distances are respectively 453.163436m, 441.958833m, 449.931227m.As shown in Figure 6.According to the collaborative air speed obtained, the average velocity that can obtain flying after path is optimized is respectively: 1.57m/s, 1.84m/s, 1.73m/s arrive path in the time of after optimization.

B. attacking angle is 0 °

When considering that three unmanned planes fly impact point from same direction, aircraft initial position, target location and to arrive from different directions situation identical.Unmanned plane arrives in turn with 0 ° of course angle target.Before optimizing, flying distance is respectively 517.221613m, 472.987332m, 489.619395m; After collaborative, flying distance is respectively: 451.109415m, 445.080861m, 454.099246m; Obtain the estimated time of arrival (ETA) t α=291s of team, three unmanned plane during flying speed are respectively 1.55m/s, 1.53m/s, 1.56m/s.

3. the simulation experiment result

The result that ant colony algorithm is applied to trajectory planning is carried out emulation, and result shows that ant colony algorithm can complete the planning of collaborative flight path well; The flight track cost is as collaborative function, and every unmanned plane arrives the time range of task object, utilizes collaborative function to pass on.Obtain allowing multiple UAVs arrive task object simultaneously, while team's Least-cost, and obtain single frame unmanned plane cost suboptimum flight path as far as possible.And the direction of attack at impact point place is processed.

Claims (2)

1. the multiple no-manned plane path planning method based on the collaborative algorithm of looking for food of bee colony, is characterized in that, comprises the steps:

(1), according to environmental model, the details of initialization trajectory planning task, carry out the coordinate discrete transform;

(2) the collaborative algorithm parameter of looking for food of initialization bee colony, comprise threshold value T search time, and iterations is K, Bas=0 search time of initialization algorithm, and initial iterations K=1 population total N, lead the machine number N ₁with follow the machine number N ₂, N=N ₁+ N ₂;

(3), according to the unmanned aerial vehicle flight path parameter, calculate the cost f in the flight path path of every unmanned plane ^j(j=1,2,3..., n),

x (j) is j bar air route, J _i(x (j)) is i item routeing target penalty function in the routeing target, ω _ifor the weight coefficient of each penalty, j is the unmanned aerial vehicle flight path ordinal number, and n is the flight path sum, the total item that k is needs assessment in evaluation function;

(4) every unmanned plane, according to current position calculation current path cost, if the current path cost is lower than original route cost, upgrades the position of unmanned plane: for i unmanned plane, be created in the integer j of [1, D], [1, NE] integer k, j parameter of i unmanned plane used change, the unmanned aerial vehicle flight path that calculates the new cost value after undated parameter and select a low-cost is as new flight path,

unmanned plane position vector eigenwert, j ∈ 1,2 ..., Q}, rand (0,1) is (0,1) upper equally distributed random number, x _max, x _minrespectively x _ihigher limit and lower limit;

(5) according to the cost value of every unmanned plane, the machine of following selects to lead object, and each leads machine to recruit one and follows machine, leading the machine surrounding space to continue the search new route, repeating step (4), if this new route is lower than original route cost, that is converted into the machine that leads at random, upgrade the position of every unmanned plane, continue to survey, search time, Bas set to 0 again, if this new route is higher than original route cost, the maintenance search, Bas adds 1;

(6) if Bas search time is greater than definite critical value, unmanned plane is abandoned current path, and the search new route, re-execute (2);

(7) preserve the parameter of optimal path and calculate optimal value;

(8) check whether reach the iterations upper limit, reach and finish search, otherwise repeating step (4) is to step (7).

2. a kind of multiple no-manned plane path planning method based on the collaborative algorithm of looking for food of bee colony according to claim 1 is characterized in that: the details of described initialization trajectory planning task, carry out the coordinate discrete transform, and comprise i navigation spots N _icoordinate be:

$\left\{\begin{array}{c}{x}_i=x_{i-1}+r\cos \mathrm{\θ}\cos \mathrm{\α}-r\sin \mathrm{\θ}\sin \mathrm{\α}\\ y_i=y_{i-1}+r\cos \mathrm{\θ}\sin \mathrm{\α}+r\sin \mathrm{\θ}\cos \mathrm{\α}\end{array}\right.,r\≥\frac{d(s,g)}{N},r\≥{4l}_{\min },$

S, g are respectively starting point and the terminal in initialization air route, and r is single air route section step-length, the air line distance that d (s, g) is the air route Origin And Destination, and N is the navigation spots number that air route allows at most, 4l _minfor the minimum flight air route segment length of unmanned plane constraint, (x _i, y _i) be the unmanned plane position coordinates, the angle that θ is unmanned plane route and transverse axis, α is the anglec of rotation.

Images