CN105841702A - Method for planning routes of multi-unmanned aerial vehicles based on particle swarm optimization algorithm - Google Patents

Abstract

The invention provides a method for planning the routes of multi-unmanned aerial vehicles based on a particle swarm optimization algorithm. The method comprises the following steps: establishing a three-dimensional map for planning space of the routes of the multi-unmanned aerial vehicles at first; then constructing multi-unmanned aerial vehicle route planning models under the three-dimensional map, wherein the models mainly comprise a barrier model, a route model, an unmanned aerial vehicle state model, a constraint model and a multi-unmanned aerial vehicle route planning mathematic model; and solving the problem of multi-unmanned aerial vehicle route planning under the three-dimensional map by using the particle swarm optimization algorithm. The method provided by the invention improves multi-unmanned aerial vehicle route planning capacity in a complex environment and provides technical support for air traffic management platforms for unmanned aerial vehicles, autonomous flight systems for multi-unmanned aerial vehicles, etc.

Description

A kind of multiple no-manned plane Route planner based on particle swarm optimization algorithm

Technical field

The present invention relates to a kind of multiple no-manned plane Route planner based on particle swarm optimization algorithm, belong to unmanned plane air route Planning field.

Background technology

At present, unmanned plane has been succeeded in multiple fields such as electric power, communication, meteorology, monitorings application, achieves good Economic benefit and social effect.Due to the restriction of single rack unmanned plane self software and hardware condition, it is difficult to competent day by day complicated answering By environment and diversified mission requirements.It is the important of following Development of UAV application that multiple no-manned plane has worked in coordination with the pattern of task Trend, is an up unmanned plane tasks carrying efficiency, expands new task state, improves the effective way of system reliability.The U.S. SAB of air force the most once pointed out, unmanned plane should work rather than independent action in the way of a group of planes.

How man-machine the most collaborative multiple no-manned plane routeing be to realize one of key technology that is and that manage.At present, existing Multiple no-manned plane Route planner typically uses optimum formula algorithm, mainly includes the method for exhaustion, dynamic programming, Mathematical Planning, cattle Method and gradient method etc..There is bigger limitation in reality application in these methods, as lacked multiple no-manned plane routeing Effective decision-making, i.e. considers deficiency to the conflicting of information observation, the redundancy of dependency between multiple no-manned plane；The power of weighting algorithm The distribution of value, with the biggest subjectivity, lacks effective, practical weight distribution method etc..Therefore, in the urgent need to more effectively, Practical multiple no-manned plane Route planner.

Summary of the invention

Present invention aim to address the problems referred to above existing for existing multiple no-manned plane Route planner, propose one Multiple no-manned plane Route planner based on particle swarm optimization algorithm.

Technical scheme mainly comprises the steps that

Step 1: set up the three-dimensional map in multiple no-manned plane routeing space.Unmanned plane during flying space is entered with data mode Row storage, the institute being denoted as in planning space a little (x, y, set z) (x, y, z) | X_min≤x≤X_max, Y_min≤y≤ Y_max, Z_min≤z≤Z_max, wherein (x, y) represents the horizontal level of this point, and z is altitude data.Planning space after discretization is adopted Digital terrain elevation data is preserved by the form of grid.

Step 2: set up multiple no-manned plane routeing model under three-dimensional map.Mainly comprise the steps that

Step 2.1, set up barrier model.The obstacle of unmanned plane is divided into soft barrier and sound-hard obstacle.The softest barrier The description hindering thing includes: the center of barrier, the operating radius of barrier, the barrier damage probability to unmanned plane.Firmly Barrier is typically due to specific factor and limits the region that can not leap, once pass through, will damage.Sound-hard obstacle can be seen Becoming the special circumstances of soft barrier, its description is the same with the description of soft barrier, the simply damage probability value 0 or 1 of obstacle.Always Damage probability ω equal to each parallel damage probability ω_iSuperposition, and each parallel damage probability ω_iCalculating comprise with it Serial damage probability ω_ijMeet following relation:

$\left\{\begin{array}{c}\mathrm{\ω}=\underset{i=1}{\overset{n}{\Σ}}{\mathrm{\ω}}_i\\ {\mathrm{\ω}}_i=1-\underset{j=1}{\overset{m}{\Σ}}(1-{\mathrm{\ω}}_{ij})\end{array}\right.$

Step 2.2, set up path model.For every frame unmanned plane, feasible path can be regarded as one from starting point to end Point, the broken line being made up of some line segments, can represent by the end points sequence of these broken lines.For avoiding with the angle being excessively sharp Path.Use the processing method of " correction ": addition one end circular arc at sharp corner, carry out correspondence seamlessly transits process.This The radius of individual circular arc is chosen as the min. turning radius of unmanned plane, meets the mobility constraint of unmanned plane, and ensures to constitute Between this circular arc and two lines even tangent, replace original corner point with two point of contacts obtained.

Step 2.3, set up the state model of unmanned plane.This patent is according to the practical operation situation of unmanned plane, by unmanned plane Divide into: ready, work, make a return voyage, 4 states out of control.

Step 2.4, set up Path Planning for Unmanned Aircraft Vehicle major constraints model.This patent selects unmanned plane vertical direction maximum to turn Bent angle constraint, the constraint of horizontal direction maximum turning angle, maximum radius of turn constraint, the constraint of farthest flying distance, flying height are about Bundle, special way point etc. are as the constraints carrying out routeing.Specific as follows:

(1) vertical direction maximum turning angle constraint

$arctan\left(\frac{|z_{i+1}-z_i|}{\sqrt{{(x_{i+1}-x_i)}^2+{(y_{i+1}-y_i)}^2}}\right)\≤{\mathrm{\θ}}_{max}$

In formula, i represents to be currently i-th section of flight path, (x_i, y_i, z_i) and (x_i+1, y_i+1, z_i+1) represent current way point respectively With the way point to be selected position coordinates in planning space, θ_maxRepresent unmanned plane maximum angle of turn in vertical direction.

(2) horizontal direction maximum turning angle constraint

In formula,Represent unmanned plane maximum angle of turn in the horizontal direction.

(3) min. turning radius constraint

R_i≥R_min

In formula, R_iRadius of turn when i & lt is turned, R is carried out for planning flight path_minMaximum radius of turn for unmanned plane. R_minCalculated by following formula:

$R_{min}=V_{min}^2/g\×\sqrt{n_{ymax}^2-1}$

In formula, V_minFor the minimum flying speed of unmanned plane, n_{y max}Maximum normal g-load for unmanned plane.

(4) farthest flying distance constraint

$\underset{i}{\mathrm{\Σ}}\left|\right|l_i\left|\right|\≤L_{max}$

In formula, l_iRepresent the flying distance of i-th section of flight path, L_maxFor allowing farthest flying distance.

(5) flying height constraint

H_min≤H_i≤H_max

In formula, H_iFor current flight height, H_minHeight, H is flown for minimum_maxHeight is flown for the highest.

(6) special way point

Special way point includes the charging for distributing on air route or the post house changing battery for unmanned plane, is used for improving nothing Man-machine flying power.This type of point is regarded as special way point treat, when, after the too low warning of unmanned plane electricity, selecting corresponding nearby Special way point, to unmanned plane battery charging or change.For just leaving the unmanned plane of special way point, to it Carry out during routeing that it is motor-driven and flying power is considered by peak.

Step 2.5, set up multiple no-manned plane routeing mathematical model.The trajectory planning problem of multiple no-manned plane is planning space Inside meet particular requirement, and flight Least-cost, the set of a series of flight path nodes from flight starting point to the point of flight eventually, It is expressed as

$\left\{\begin{array}{c}\min C\left(p\right)\\ s.t.g\left(P\right)\≥0\end{array}\right.$

In formula, C (p) by the cost function of planning unmanned plane during flying air route p, g (p) is constraints.

Step 3: use particle swarm optimization algorithm multiple no-manned plane routeing.Specifically include following steps:

Step 3.1, according to set up three-dimensional map under multiple no-manned plane routeing model, choose particle swarm optimization algorithm Decision variable, and determine the bound of decision variable；

Step 3.2: according to multiple no-manned plane routeing model under the three-dimensional map set up, particle swarm optimization algorithm is set Object function.

Step 3.3: population number in particle cluster algorithm iterative process, maximum iteration time, particle maximum flight speed are set Degree, Studying factors, inertia weight scheduling algorithm basic parameter.

Step 3.4: under different initial condition, obtains the routeing result meeting constraint requirements by algorithm iteration.

Present invention is characterized in that

1. the multiple no-manned plane Route planner that the present invention provides realizes based on Three-dimensional Numeric Map, is not only restricted to the most unmanned The task context of machine and working environment.By setting up different Three-dimensional Numeric Map, this method application can be realized easily The extension of scene.

2. the present invention uses particle swarm optimization algorithm multiple no-manned plane routeing problem.Particle swarm optimization algorithm has Do not rely on degree of the passing information of problem, need the parameter adjusted few, convergence precision high；Meanwhile, algorithm can be located simultaneously Different constraints in reason multiple no-manned plane routeing and some particular/special requirement.

3. the present invention is when multiple no-manned plane routeing models, to some specific functions in aerial pipe platform (as given Unmanned plane charging, the post house of replacing battery, hesitation point etc.) embodied.Therefore, the multiple no-manned plane air route that the present invention proposes Planing method can be applicable among following unmanned plane aerial pipe platform.

Accompanying drawing explanation

Fig. 1 is the flowchart of the present invention.

Detailed description of the invention

In order to make those skilled in the art be more fully understood that the application scheme, below in conjunction with the accompanying drawing 1 in the application, Technic relization scheme in the application is illustrated.

Step 1: set up the three-dimensional map in multiple no-manned plane routeing space.Unmanned plane during flying space is entered with data mode Row storage, the institute being denoted as in planning space a little (x, y, set z) (x, y, z) | X_min≤x≤X_max, Y_min≤y≤ Y_max, Z_min≤z≤Z_max, wherein (x, y) represents the horizontal level of this point, and z is altitude data.Planning space after discretization is adopted Digital terrain elevation data is preserved by the form of grid.

Step 2: set up multiple no-manned plane routeing model under three-dimensional map.Mainly include barrier model, path model, Unmanned plane state model, restricted model and multiple no-manned plane routeing mathematical model.Wherein restricted model includes again vertical direction Maximum turning angle constraint, the constraint of horizontal direction maximum turning angle, min. turning radius constraint, farthest flight-follow constraint, flight Highly constrained and special way point etc..

Step 3: use particle swarm optimization algorithm multiple no-manned plane routeing.Specifically include selection Particle Swarm Optimization Method decision variable also determines its bound, arranges the object function of algorithm, arrange population number in particle cluster algorithm iterative process, Maximum iteration time, particle maximum flying speed, Studying factors, inertia weight scheduling algorithm basic parameter.Navigate for multiple no-manned plane Circuit planning problem, the reference of particle swarm optimization algorithm major parameter arranges value and is: population population number 20～50, and algorithm is maximum Iterations 50～100, particle maximum flying speed is the 1/10～1/5 of relevant variable hunting zone, and Studying factors is 2, used The initial value of property weight is 0.9, and stop value is 0.4.Finally, the routeing knot of constraint requirements is met by algorithm iteration acquisition Really.

Claims (1)

1. a multiple no-manned plane Route planner based on particle swarm optimization algorithm, mainly comprises the steps that

Step 1: set up the three-dimensional map in multiple no-manned plane routeing space.Unmanned plane during flying space is deposited with data mode Storage, the institute being denoted as in planning space a little (x, y, set z) (x, y, z) | X_min≤x≤X_max, Y_min≤y≤Y_max, Z_min≤z≤Z_max, wherein (x, y) represents the horizontal level of this point, and z is altitude data.Planning space after discretization uses grid The form of lattice preserves digital terrain elevation data.

Step 2: set up multiple no-manned plane routeing model under three-dimensional map.Mainly comprise the steps that

Step 2.1, set up barrier model.The obstacle of unmanned plane is divided into soft barrier and sound-hard obstacle.The softest barrier Description include: the center of barrier, the operating radius of barrier, the barrier damage probability to unmanned plane.Hard obstacle Thing is typically due to specific factor and limits the region that can not leap, once pass through, will damage.Sound-hard obstacle can regard soft as The special circumstances of barrier, its description is the same with the description of soft barrier, the simply damage probability value 0 or 1 of obstacle.Total damage Hinder probability ω equal to each parallel damage probability ω_iSuperposition, and each parallel damage probability ω_iCalculating comprise serial with it Damage probability ω_ijMeet following relation:

$\left\{\begin{array}{c}\mathrm{\ω}=\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{\mathrm{\ω}}_i\\ {\mathrm{\ω}}_i=1-\underset{j=1}{\overset{m}{\Σ}}(1-{\mathrm{\ω}}_{ij})\end{array}\right.$

Step 2.2, set up path model.For every frame unmanned plane, feasible path can be regarded as one from origin-to-destination, by The broken line of some line segments composition, can represent by the end points sequence of these broken lines.For avoiding the road with the angle being excessively sharp Footpath,.Use the processing method of " correction ": addition one end circular arc at sharp corner, carry out correspondence seamlessly transits process.This The radius of circular arc is chosen as the min. turning radius of unmanned plane, meets the mobility constraint of unmanned plane, and ensures that composition should Between circular arc and two lines even tangent, replace original corner point with two point of contacts obtained

Step 2.3, set up the state model of unmanned plane.Unmanned plane, according to the practical operation situation of unmanned plane, is distinguished by this patent For: ready, work, make a return voyage, 4 states out of control.

Step 2.4, set up Path Planning for Unmanned Aircraft Vehicle major constraints model.This patent selects unmanned plane vertical direction maximum turning angle Retrain, horizontal direction maximum turning angle retrains, maximum radius of turn retrains, farthest flying distance retrains, flying height retrains, spy Different way points etc. are as the constraints carrying out routeing.Specific as follows:

(1) vertical direction maximum turning angle constraint

$\arctan \left(\frac{|z_{i+1}-z_i|}{\sqrt{{(x_{i+1}-x_i)}^2+{(y_{i+1}-y_i)}^2}}\right)\≤{\mathrm{\θ}}_{max}$

In formula, i represents to be currently i-th section of flight path, (x_i, y_i, z_i) and (x_i+1, y_i+1, z_i+1) represent current way point respectively and treat Select way point position coordinates in planning space, θ_maxRepresent unmanned plane maximum angle of turn in vertical direction.

(2) horizontal direction maximum turning angle constraint

In formula,Represent unmanned plane maximum angle of turn in the horizontal direction.

(3) min. turning radius constraint

R_i≥R_min

In formula, R_iRadius of turn when i & lt is turned, R is carried out for planning flight path_minMaximum radius of turn for unmanned plane.R_minBy Following formula calculates:

$R_{min}=V_{min}^2/g\×\sqrt{n_{ymax}^2-1}$

In formula, V_minFor the minimum flying speed of unmanned plane, n_ymaxMaximum normal g-load for unmanned plane.

(4) farthest flying distance constraint

$\underset{i}{\Σ}\left|\right|l_i\left|\right|\≤L_{max}$

In formula, l_iRepresent the flying distance of i-th section of flight path, L_maxFor allowing farthest flying distance.

(5) flying height constraint

H_min≤H_i≤H_max

In formula, H_iFor current flight height, H_minHeight, H is flown for minimum_maxHeight is flown for the highest.

(6) special way point

Special way point includes the charging for distributing on air route or the post house changing battery for unmanned plane, is used for improving unmanned plane Flying power.This type of point is regarded as special way point treat, when, after the too low warning of unmanned plane electricity, selecting corresponding special nearby Different way point, to the charging of unmanned plane battery or replacing.For just leaving the unmanned plane of special way point, it is being carried out During routeing, it is motor-driven and flying power is pressed peak and considered.

Step 2.5, set up multiple no-manned plane routeing mathematical model.The trajectory planning problem of multiple no-manned plane is full in planning space Foot particular requirement, and flight Least-cost, the set of a series of flight path nodes from flight starting point to the point of flight eventually, represent For

$\left\{\begin{array}{c}\min C\left(p\right)\\ s.t.g\left(P\right)\≥0\end{array}\right.$

In formula, C (p) by the cost function of planning unmanned plane during flying air route p, g (p) is constraints.

Step 3: use particle swarm optimization algorithm multiple no-manned plane routeing.Specifically include following steps:

Step 3.1, according to set up three-dimensional map under multiple no-manned plane routeing model, choose the decision-making of particle swarm optimization algorithm Variable, and determine the bound of decision variable；

Step 3.2: according to multiple no-manned plane routeing model under the three-dimensional map set up, the target of particle swarm optimization algorithm is set Function.

Step 3.3: population number in particle cluster algorithm iterative process, maximum iteration time, particle maximum flying speed, are set Practise the factor, inertia weight scheduling algorithm basic parameter.

Step 3.4: under different initial condition, obtains the routeing result meeting constraint requirements by algorithm iteration.