CN109445463B - Unmanned aerial vehicle dynamic route planning method - Google Patents

Abstract

The invention relates to a method for planning a dynamic route of an unmanned aerial vehicle, which comprises the following steps: 1) establishing a route cost estimation model: constructing a communication signal threat model and a terrain surface threat model, weighting and superposing the communication signal threat model and the terrain surface threat model to form a comprehensive threat model, and realizing the evaluation of a planned route on the basis of the comprehensive threat model; 2) calculating a flight route: and constructing a comprehensive coordinate set random flight line set, realizing dynamic updating of the flight line set aiming at the threat change condition, optimizing a planned flight path point set in the flight line, and selecting the flight line with the maximum safety performance and the capability of flying for each airplane. The invention relates to a method for selecting routes in consideration of flight safety, range, airplane performance and the like. The calculation results show that the planned flight route realizes real-time planning and route optimization of the global optimal route in a complex environment, and can avoid various dynamic threats in time, thereby improving the safety and reliability of flight.

Description

Unmanned aerial vehicle dynamic route planning method

Technical Field

The invention belongs to the field of unmanned aerial vehicle route planning, and relates to a method for planning an optimal flight route by using information such as unmanned aerial vehicle performance, environmental threat and the like, in particular to a dynamic route planning method for an unmanned aerial vehicle.

Background

So far, the research on the air route planning in complex environment at home and abroad is still preliminary, and a complete and systematic unmanned aerial vehicle dynamic flight air route planning model and method are not formed yet. The unmanned aerial vehicle dynamic route planning in the complex geographic environment is not well solved at present, how to plan the flight route with the maximum safety performance by applying an artificial clustering algorithm is a problem to be solved urgently, and the solution of the problem has extremely important application value for military use and civil use no matter how to avoid dynamic threats and realize aviation safety rescue.

Disclosure of Invention

The invention aims to provide a dynamic route planning method for an unmanned aerial vehicle, which can meet the requirement that a plurality of unmanned aerial vehicles plan routes in a concealed environment and improve the flight safety and efficiency as much as possible. The method comprises the following specific steps:

1) establishing a route cost estimation model:

(1) construction of communication signal threat model f₁：

Wherein: n is_xt＝(x_i+x_i+1)/2

N in the formula (1) is the total number of track points in the route, x_n＝X_e，y_n＝Y_e，(x_i,y_i)、(nx_i,ny_i) And (X)_e,Y_e) Respectively representing the ith track point coordinate, the ith adjusted new track point coordinate and the target point coordinate, wherein tn is the total number of communication threats, (xc)_k,yc_k) And R_kRespectively representing the center coordinate and the influence radius of the kth communication threat;

(2) construction of a terrain surface threat model f₂：

In the formula (2), Mn represents the number of peaks, Hmax_i(xo) in terms of the height of the peak of the ith peak_j,yo_j) And (x)_sj，y_sj) Respectively representing the plane coordinate of the peak top of the jth peak and parameters descending along the x direction and the y direction;

(3) threat model f by communication signal₁And a terrain surface threat model f₂And weighted superposition to form a comprehensive threat model F, wherein the mathematical description is as follows:

wherein p is₁、p₂Weights, f, corresponding to communication threats and terrain threats, respectively_i1、f_i2Respectively a communication threat value and a terrain threat value at the ith track point;

(4) flight line evaluation function fit:

2) calculating a flight route, and specifically comprising the following steps:

(1) constructing a comprehensive coordinate set (xy) random route set

xy＝(lb+(ub-lb)*rand(Nfp)*100) (5)

xy＝x*100+y (6)

xy＝{xy₁，xy₂，...，xy_i，...，xy_Nfp} (7)

In formulas (5) - (7), xy represents a comprehensive track point coordinate set, (x, y) are coordinate values of track points, rand (Nfp) generates Nfp random numbers between 0 and 1, lb and ub respectively represent upper limit and lower limit values of the track point coordinates, and xy and ub respectively represent upper limit and lower limit values of the track point coordinates_iThe comprehensive coordinate value of the ith track point is represented;

(2) dynamic update of a set of routes

ff₁＝xy_i+c₁*rand*(bestXY-xy_i)+c₂*rand*(pxy_i-xy_i) (9)

P, V in formula (8)_pRespectively representing a randomly generated probability value and a probability value threshold, ff₁、ff₂Respectively is a comprehensive coordinate value corresponding to the random probability value less than or equal to the probability threshold value; c in formulae (9) to (10)₁And c₂Respectively representing global and local route adjustment parameters, rand representing generated random numbers between 0 and 1, bestXY and pxy_iRespectively representing coordinate values corresponding to the global and ith integrated track points, mean P representing the average integrated track point set, a₁And a₂Representing the coefficients, pFit, corresponding to means P and pxy, respectively_iAnd pFitn respectively, corresponds to pxy_iAnd the cost value of the global optimal route, sumFit represents the sum of the cost of the global optimal route, d is a parameter for avoiding the route from having an infeasible solution, and realmin represents the minimum value which can be represented by a floating point number type in Matlab software;

(3) for planned intra-route track point set { (x)₁，y₁，h₁),(x₂，y₂，h₂)，…(x₁₂，y₁₂，h₁₂) Optimization is carried out, and the formula is as follows:

b0＝1.0/6*(-1*u³+3*u²-3*u+1) (11)

b1＝1.0/6*(3*u³-6*u²+4) (12)

b2＝1.0/6*(-3*u³+3*u²+3*u+1) (13)

b3＝1.0/6*u³ (14)

x_n＝b0*x(i)+b1*x(i+1)+b2*x(i+2)+b3*x(i+3) (15)

y_n＝b0*y(i)+b1*y(i+1)+b2*y(i+2)+b3*y(i+3) (16)

h_n＝b0*h(i)+b1*h(i+1)+b2*h(i+2)+b3*h(i+3) (17)

in the formulae (11) to (17), h (i) represents the height corresponding to the ith track pointThe spline curve parameters are u, b0, b1, b2 and b3 are coefficients of basis functions, x_n、y_nAnd h_nThe abscissa, ordinate and elevation of the corresponding curve.

Compared with the prior art, the invention has the advantages that: the real-time planning and route optimization of the global optimal route in the complex environment are realized, and various dynamic threats can be avoided in time.

Detailed Description

The technical solution of the present invention is further illustrated by the following examples. In the embodiment, 1 winged dragon unmanned aerial vehicle executes the task of disaster monitoring, and 3 terrain threats are totally arranged at positions of (55km ), (15km, 25km) and (10km, 15 km); a total of 7 communication threats at (55km ), (10km, 50km), (15km, 25km), (30km, 25km), (10km, 15km), (40km, 15km) and (40km, 10 km);

the parameters are set as follows: the number of the sets of the to-be-selected routes is 30, the number of the tracks is 12, the global mean P and the adjustment parameter c of the local route pxy₁、c₂1.5 and 1.5 respectively, and averaging the coefficient a of the integrated track point set mean P₁And coefficient a of the set of synthetic track points pxy for the optimal route₂Are all 1; the coordinates of the departure point and the target point are (1km, 1km, 150m) and (60km, 60km, 150m), respectively. The method comprises the following specific steps:

1) establishing a route cost estimation model:

(1) construction of communication signal threat model f₁：

nx_i＝(x_i+x_i+1)/2

(2) Construction of a terrain surface threat model f₂：

f₂(x_i,y_i)＝100*exp(-(x_i-55)²/5²-(y_i-55)²/5²)+100*exp(-(x_i-10)²/5²-(y_i-50)²/5²)+100*exp(-(x_i-15)²/7²-(y_i-25)²/6²)

(2)

(3) Threat model f by communication signal₁And a terrain surface threat model f₂And weighted superposition to form a comprehensive threat model F, wherein the mathematical description is as follows:

(4) flight line evaluation function fit:

2) calculating a flight route, and specifically comprising the following steps:

(1) constructing a comprehensive coordinate set (xy) random route set:

lb＝[1，1，1，1，1，1，1，1，1，1，1，1]

ub＝[60，60，60，60，60，60，60，60，60，60，60，60]

xy＝(lb+(ub-lb)*rand(1,12)*100) (5)

xy＝x*100+y (6)

xy＝{1001，2002，…，6060} (7)

(2) dynamic update of a set of routes

prob＝rand(30，1)*0.2+0.8

ff₁＝xy_i+1.5*rand*(bestX-xy_i)+1.5*rand*(pX(i,：)-xy_i) (9)

ff₂＝xy_i+rand*(meanP-xy_i)*1*exp(-pFit(i)/(sumPfit+realmin)*30))+1*(rand*2-1)*(pX(person,：)-xy_i)*exp(-pFit(person)-pFit(i))/(|pFit(person)-pFit(i)|+realmin)*pFit(person)/(sumPfit+realmin)*30)(10)

(3) Optimizing a set of planned intra-route track points { (1, 1, 150), … (60, 60, 150) }:

u＝0∶0.01∶1

b0＝1.0/6*(-1*u³+3*u²-3*u+1) (11)

b1＝1.0/6*(3*u³-6*u²+4) (12)

b2＝1.0/6*(-3*u³+3*u²+3*u+1) (13)

b3＝1.0/6*u³ (14)

x_n＝b0*x(i)+b1*x(i+1)+b2*x(i+2)+b3*x(i+3) (15)

y_n＝b0*y(i)+b1*y(i+1)+b2*y(i+2)+b3*y(i+3) (16)

h_n＝b0*h(i)+b1*h(i+1)+b2*h(i+2)+b3*h(i+3) (17)

and finishing the route planning.

Table 1 shows example communication and terrain threat parameters

Table 2 shows the results of the calculation of the dynamic flight path and the evaluation of the total flight path of the unmanned aerial vehicle

Experimental number	f1	f2	Total voyage evaluation value (kilometer)	Planning time (seconds)
					1	0.5	0.5	96.04	86.05

Table 2 shows that the weights of the landform and communication information threats in the experiment are both 0.5, the planned flight distance is 96.04 km, which is far less than the maximum flight distance of 4000 km, and the maximum flight speed is 280 km/h, and the flight time is at least 64.29 seconds and far longer than the real-time planning time of 5 seconds under the limitation of the minimum track segment of 5km, so that the online flight route planning can be realized, and the safety guarantee is provided for emergency rescue.

Claims (1)

1. An unmanned aerial vehicle dynamic route planning method is characterized by comprising the following steps:

1) establishing a route cost estimation model:

(1) construction of communication signal threat model f₁：

Wherein: nx_i＝(x_i+x_i+1)/2

N in the formula (1) is the total number of track points in the route, x_n＝X_e，y_n＝Y_e，(x_i,y_i)、(nx_i,ny_i) And (X)_e,Y_e) Respectively representing the ith track point coordinate, the ith adjusted new track point coordinate and the target point coordinate, wherein tn is the total number of communication threats, (xc)_k,yc_k) And R_kRespectively representing the center coordinate and the influence radius of the kth communication threat;

(2) construction of a terrain surface threat model f₂：

In the formula (2), Mn represents the number of peaks, Hmax_i(x) is the elevation of the peak top of the ith peak_oj,y_oj) And (x)_sj,y_sj) Respectively representing the plane coordinate of the peak top of the jth peak and parameters descending along the x direction and the y direction;

(3) threat model f by communication signal₁And a terrain surface threat model f₂And weighted superposition to form a comprehensive threat model F, wherein the mathematical description is as follows:

wherein p is₁、p₂Weights, f, corresponding to communication threats and terrain threats, respectively_i1、f_i2Respectively a communication threat value and a terrain threat value at the ith track point;

(4) flight line evaluation function fit:

2) calculating a flight route, and specifically comprising the following steps:

(1) constructing a comprehensive coordinate set (xy) random route set

xy＝(lb+(ub-lb)*rand(Nfp)*100) (5)

xy＝x*100+y (6)

xy＝{xy₁，xy₂，...，xy_i，...，xy_Nfp} (7)

In formulas (5) - (7), xy represents a comprehensive track point coordinate set, (x, y) are coordinate values of track points, rand (Nfp) generates Nfp random numbers between 0 and 1, lb and ub respectively represent upper limit and lower limit values of the track point coordinates, and xy and ub respectively represent upper limit and lower limit values of the track point coordinates_iThe comprehensive coordinate value of the ith track point is represented;

(2) dynamic update of a set of routes

ff₁＝xy_i+c₁*rand*(bestXY-xy_i)+c₂*rand*(pxy_i-xy_i) (9)

P, V in formula (8)_pRespectively representing a randomly generated probability value and a probability value threshold, ff₁、ff₂Respectively is a comprehensive coordinate value corresponding to the random probability value less than or equal to the probability threshold value; c in formulae (9) to (10)₁And c₂Respectively representing global and local route adjustment parameters, rand representing generated random numbers between 0 and 1, bestXY and pxy_iRespectively representing coordinate values corresponding to the global and ith integrated track points, mean P representing the average integrated track point set, a₁And a₂Representing the coefficients, pFit, corresponding to means P and pxy, respectively_iAnd pFitn respectively, corresponds to pxy_iAnd the cost value of the global optimal route, sumFit represents the sum of the cost of the global optimal route, d is a parameter for avoiding the route from having an infeasible solution, and realmin represents the minimum value which can be represented by a floating point number type in Matlab software;

(3) for planned intra-route track point set { (x)₁，y₁，h₁),(x₂，y₂，h₂)，…(x₁₂，y₁₂，h₁₂) Optimization is carried out, and the formula is as follows:

b0＝1.0/6*(-1*u³+3*u²-3*u+1) (11)

b1＝1.0/6*(3*u³-6*u²+4) (12)

b2＝1.0/6*(-3*u³+3*u²+3*u+1) (13)

b3＝1.0/6*u³ (14)

x_n＝b0*x(i)+b1*x(i+1)+b2*x(i+2)+b3*x(i+3) (15)

y_n＝b0*y(i)+b1*y(i+1)+b2*y(i+2)+b3*y(i+3) (16)

h_n＝b0*h(i)+b1*h(i+1)+b2*h(i+2)+b3*h(i+3) (17)

h (i) in the formulas (11) to (17) represents the elevation corresponding to the ith track point, the spline curve parameter is u, b0, b1, b2 and b3 are coefficients of basis functions, and x_n、y_nAnd h_nThe abscissa, ordinate and elevation of the corresponding curve.