CN104155999A - Time-sensitive task dynamic allocation algorithm in battlefield environment for multiple unmanned aerial vehicles - Google Patents

Abstract

The invention discloses a time-sensitive task dynamic allocation algorithm in the battlefield environment for multiple unmanned aerial vehicles. The algorithm can realize that multiple unmanned aerial vehicles strike multiple time-sensitive targets in the battlefield environment; according to the algorithm, a task auction idea is used to construct a time-sensitive task dynamic allocation model, and the task execution time, damage ability and strike earnings as well as the influences of the execution of the current task on time-sensitive task time window width and threat degree are mainly considered by the model; the strike path planning of the unmanned aerial vehicles is realized by adopting a model predictive control algorithm, and that the strike paths of unmanned aerial vehicles are optimized under the condition that the unmanned aerial vehicles keep away from threat regions is mainly completed.

Description

Quick task dynamic allocation algorithm during multiple no-manned plane under battlefield surroundings

Technical field

The present invention relates to a kind of allocation algorithm, quick task dynamic allocation algorithm while relating in particular under a kind of battlefield surroundings multiple no-manned plane.

Background technology

Quick target (Time-sensitive Targets, TST) when time-sensitive type target is called for short, the target that mainly finger must complete discovery, location, identification and hit in limited " time window " or " belligerent chance ".The very strong uncertainty of status flag existence due to TST, make unmanned plane (Unmanned Aerial Vehicle, UAV) form into columns when hitting the type target, need to be according to time window width and the current state of TST, the performance of UAVs and useful load difference, carry out in time task and hit the dynamic optimization in path, to guarantee the form into columns maximization of quick task income while carrying out of UAVs.

At present, the coupling that researchist executes the task for UAVs and time-sensitive feature, the isomerism of UAVs system and dispersing character, many task dynamic allocation algorithm have been proposed, as: distributed task scheduling consistency algorithm, task dynamic programming algorithm, the task allocation algorithms of optimizing based on fitness, task allocation algorithms based on market auction with based on game theoretic task allocation algorithms etc.Wherein, the binding consistency algorithm (Consensus-based Bundle Algorithm, CBBA) being proposed by people such as Massachusetts Institute Technology aviation control laboratory Jonathan How, Han-LimChoi has been subject to extensive concern.This algorithm combines combination auction mechanism and distributed task scheduling congruity theory, can under the condition of task coupling, realize isomery multiple agent combined task and distribute.CCBBA (Coupled-Constraint Consensus-Based Bundle Algorithm), on the basis of CBBA, is further developed in this laboratory.This algorithm has been introduced intelligent body and has been executed the task the time, quick task distribution capability when CBBA has been possessed.Said method is all under the fixing assumed condition of TST time window, utilizes evolution algorithm or greedy search algorithm to obtain optimum solution, lacks task dynamic assignment ability in the time dependent situation of TST time window, and this will affect the effect of UAVs to TST strike.

Summary of the invention

For the problems referred to above, quick task dynamic allocation algorithm when the present invention proposes a kind of multiple no-manned plane, realizes the strike of the quick target when a plurality of of multiple UAVs under battlefield surroundings.Quick task dynamic assignment model when this algorithm utilizes task auction thought to build, this model mainly considered unmanned plane task execution time, injure ability, hit income, and carry out current task when follow-up quick task time window width and the impact of threaten degree; Unmanned plane hits path planning and adopts Model Predictive Control Algorithm to realize, and mainly completes the strike path optimization of unmanned plane under threatening area is evaded.

For reaching above-mentioned technical purpose, quick task dynamic allocation algorithm when the present invention has adopted under a kind of battlefield surroundings multiple no-manned plane, UAVs forms into columns and adopts Leader-Follower structure to carry out TST strike task, Leader_UAV mainly bears decision maker's role, and Follower_UAV bears striker's role.Leader_UAV will according to time quick job change situation, and Follower_UAV execute the task the time, injure ability and hit the factors such as income and complete formation task and distribute.On the basis of distributing in Leader_UAV task, Follower_UAV will complete flight path planning and TST will be hit according to task definition, with Leader_UAV, serve as task auctioneer, and Follower_UAV serves as bidder.Follower_UAV by auction price matrix cost_auction separately (comprising: and distance d between TST, carry out time t that TST hits task, to the ability of the injuring Damage of TST and hit income VLR) send to Leader_UAV.On this basis, Leader_UAV, under the constraint of TST time window width duration, selects bidder according to cost_auction, the distribution of finishing the work.Duration is subject to Follower_UAV to carry out the impact of current task, and its expression formula is suc as formula shown in (1).Each element expression in cost_auction is suc as formula shown in (2).

duration _j(k)＝duration _j(k-1)+variation _ij (1)

In formula, duration _jrepresent TST _jtime window width, variation _ijrepresent

Follower_UAV _icarry out current task to TST _jthe impact of time window width.

cost_auction _ij(k)＝d _ij(k)/max(d _ij(k))+Damage _ij(k)

+VLR _ij(k)+max(t _ij(k))/t _ij(k) (2)

+max(P_threat _ij(k))/P_threat _ij(k)

Suppose V={1,2 ..., N} is Follower_UAV quantity, in forming into columns, the quantity of UAVs is N+1, and T={1,2 ..., M} is TST quantity.In formula, d _ijrepresent Follower_UAV _i(i ∈ V) and TST _jdistance between (j ∈ T); t _ijrepresent Follower_UAV _icarry out TST _jhit task time; Damage _ijrepresent Follower_UAV _ito TST _jthe ability of injuring; VLR _ijrepresent Follower_UAV _ihit TST _jbenefit, it is Damage _ijwith bullet-loading capacity Load _ifunction, its expression formula is suc as formula shown in (3); P_threat _ijrepresent TST _jto Follower_UAV _ithreat probabilities, will pass through TST _jradar to Follower_UAV _iinstantaneous tracking probability P _dijat cumulative time section T _son integration obtain, its expression formula is suc as formula shown in (4).

${\mathrm{VLR}}_{\mathrm{ij}}\left(k\right)={\mathrm{VLR}}_{\mathrm{ij}}(k-1)\·\frac{{\mathrm{Load}}_i\left(k\right)-{\mathrm{Load}}_i(k-1)}{{\mathrm{Damage}}_{\mathrm{ij}}\left(k\right)-{\mathrm{Damage}}_{\mathrm{ij}}(k-1)}---\left(3\right)$

In formula, Load _iand Load (k) _i(k-1) represent respectively Follow_UAV _ithe bullet-loading capacity of current time and previous moment.

$P\_{\mathrm{threat}}_{\mathrm{ij}}\left(k\right)=\underset{k=c\_t-T_s}{\overset{c\_t}{\mathrm{\Σ}}}{P}_{\mathrm{dij}}\left(k\right)---\left(4\right)$

In formula, c_t is current sampling instant, T _sfor the cumulative time section of instant probability, P _dijtST _jradar to Follower_UAV _iinstantaneous tracking probability, its expression formula is suc as formula shown in (5).

$P_{\mathrm{dij}}=\left\{\begin{array}{cc}1/[1+{(c_{2j}{R_{\mathrm{ij}}}^4/{\mathrm{\&sigma;}}_i)}^{c_{1j}}]& R<R_{\max }\\ 0& R\&GreaterEqual;R_{\max }\end{array}\right.---\left(5\right)$

In formula, R _ijfollower_UAV _iwith TST _jbetween distance; c _1j, c _2jconstant, by TST _jradar type determine; σ _ifollower_UAV _iradar Cross Section (Radar Cross-Section, RCS).The RCS of UAV is the included angle X about UAV course and TST horizontal direction, the angle of pitch between TST and UAV the function of side acceleration and acceleration of gravity ratio μ, its expression formula is suc as formula shown in (6), (7), (8).

$\mathrm{\&sigma;}=\frac{\mathrm{\&pi;}\&CenterDot;a^2{b}^2{c}^2}{{(a^2{\sin }^2{\mathrm{\&lambda;}}_e{\cos }^2{\mathrm{\&mu;}}_e+b^2{\sin }^2{\mathrm{\&lambda;}}_e{\sin }^2{\mathrm{\&mu;}}_e+c^2{\cos }^2{\mathrm{\&lambda;}}_e)}^2}---\left(6\right)$

Leader_UAV is according to the cost_auction distribution of finishing the work, if Follower_UAV receives the task that Leader_UAV issues, need to calculate it and carry out the distance between TST, if this distance is less than or equal to the bee-line D_start that starts to carry out strike task, Follower_UAV starts to hit, and in original place, spirals until target is smashed; Otherwise Follower_UAV utilizes MPC algorithm to hit path planning.According to the kinematical equation of aircraft, set up nonlinear forecast model suc as formula shown in (9), (10), (11), (12), (13).

z(k+1)＝z(k)+s·sin(θ(k+1)) (11)

θ(k+1)＝θ(k)+θ ₀·u _θ(k) U _θmin≤u _θ≤U _θmax (13)

In formula, (x, y, z) represents position; S represents the air line distance between current navigation spots and next navigation spots; position angle, be constant, represent azimuthal change step, azimuthal controlled quentity controlled variable, it is the border of position angle controlled quentity controlled variable; θ is the angle of pitch, θ ₀be constant, represent the change step of the angle of pitch, u _θthe controlled quentity controlled variable of the angle of pitch, U _{θ min}, U _{θ max}it is the border of angle of pitch controlled quentity controlled variable.

u _θit is the optimum control amount obtaining by rolling optimization path cost function cost_path.The threat cost that path cost function cost_path is herein caused by state error cost, TST radar and the control cost in control domain form, and each element expression in cost_path is suc as formula shown in (14).

cost_path _ij(k)＝αc _ij ^T(k)C _ic _ij(k) (14)

+βP_threat _ij(k)+γu ^T(k)u(k)

In formula, c _ij(k) represent Follower_UAV _iwith TST _jbetween distance cost, i.e. c _ij(k)=(x _i-x _j, y _i-yj _,z _i-z _j) ^t, TST _jposition S _j(k)=(x _j, y _j, z _j) ^trepresent reference locus, α is its weights, C _ifor UAV _iand the distance cost weighting matrix between TST; P_threat _ijrepresent TST _jto Follower_UAV _ithreat ability, β is its weights; U (k) is controlled quentity controlled variable, the weights that γ is it.

The realization of quick task dynamic allocation algorithm while the invention allows for multiple no-manned plane under a kind of battlefield surroundings, comprises following steps:

Step 1:Follower_UAV submits TST observation information to Leader_UAV, and Leader_UAV produces task list, and carries out task auction according to formula (1), and the task that realizes is distributed;

Step 2:Follower_UAV is according to task allocation result, judges whether the task list of self is empty, if it is empty, and in current location waits of spiraling, until task list while be sky, is executed the task by new task list, otherwise performed step 3;

Step 3:Follower_UAV is according to the task list of self, judgement and the distance of carrying out between TST, if this is apart from being less than or equal to D_start, Follower_UAV starts to strike target, and in original place, spirals until target is smashed; Otherwise Follower_UAV carries out path planning by the optimum control amount obtaining according to formula (13), after reaching striking distance, start to strike target, and in original place, spiral until target is smashed;

Step 4:Follower_UAV carries out current task and can exert an influence to follow-up TST time window width, need to distribute and hit path and adjust UAVs formation task, will return to step 1 and carry out task weight-normality and draw, until complete the strike task of all TST.

The invention solves UAVs under battlefield surroundings to a plurality of time window dynamic changes time quick task Percussion Problems.Impact on follow-up TST time window width when this algorithm takes into full account UAVs execution current task, adopts task auction mechanism to complete strike task and dynamically updates, and realizes effective strike of TST.

Accompanying drawing explanation

Shown in Fig. 1 is quick task dynamic assignment schematic diagram when UAVs forms into columns in the present invention;

Shown in Fig. 2 is the strike path locus figure of quick task dynamic allocation algorithm while adopting in the present invention;

Embodiment

Below in conjunction with the drawings and specific embodiments, the present invention is described in more detail.

When in Fig. 1, UAVs forms into columns, quick task dynamic assignment schematic diagram is known, quick task dynamic allocation algorithm during multiple no-manned plane under a kind of battlefield surroundings, UAVs forms into columns and adopts Leader-Follower structure to carry out TST strike task, Leader_UAV mainly bears decision maker's role, and Follower_UAV bears striker's role.Leader_UAV will according to time quick job change situation, and Follower_UAV execute the task the time, injure ability and hit the factors such as income and complete formation task and distribute.On the basis of distributing in Leader_UAV task, Follower_UAV will complete flight path planning and TST will be hit according to task definition, with Leader_UAV, serve as task auctioneer, and Follower_UAV serves as bidder.Follower_UAV by auction price matrix cost_auction separately (comprising: and distance d between TST, carry out time t that TST hits task, to the ability of the injuring Damage of TST and hit income VLR) send to Leader_UAV.On this basis, Leader_UAV, under the constraint of TST time window width duration, selects bidder according to cost_auction, the distribution of finishing the work.Duration is subject to Follower_UAV to carry out the impact of current task, and its expression formula is suc as formula shown in (1).Each element expression in cost_auction is suc as formula shown in (2).

duration _j(k)＝duration _j(k-1)+variation _ij (1)

In formula, duration _jrepresent TST _jtime window width, variation _ijrepresent

Follower_UAV _icarry out current task to TST _jthe impact of time window width.

cost_auction _ij(k)＝d _ij(k)/max(d _ij(k))+Damage _ij(k)

+VLR _ij(k)+max(t _ij(k))/t _ij(k) (2)

+max(P_threat _ij(k))/P_threat _ij(k)

Suppose V={1,2 ..., N} is Follower_UAV quantity, in forming into columns, the quantity of UAVs is N+1, and T={1,2 ..., M} is TST quantity.In formula, d _ijrepresent Follower_UAV _i(i ∈ V) and TST _jdistance between (j ∈ T); t _ijrepresent Follower_UAV _icarry out TST _jhit task time; Damage _ijrepresent Follower_UAV _ito TST _jthe ability of injuring; VLR _ijrepresent Follower_UAV _ihit TST _jbenefit, it is Damage _ijwith bullet-loading capacity Load _ifunction, its expression formula is suc as formula shown in (3); P_threat _ijrepresent TST _jto Follower_UAV _ithreat probabilities, will pass through TST _jradar to Follower_UAV _iinstantaneous tracking probability P _dijat cumulative time section T _son integration obtain, its expression formula is suc as formula shown in (4).

${\mathrm{VLR}}_{\mathrm{ij}}\left(k\right)={\mathrm{VLR}}_{\mathrm{ij}}(k-1)\&CenterDot;\frac{{\mathrm{Load}}_i\left(k\right)-{\mathrm{Load}}_i(k-1)}{{\mathrm{Damage}}_{\mathrm{ij}}\left(k\right)-{\mathrm{Damage}}_{\mathrm{ij}}(k-1)}---\left(3\right)$

In formula, Load _iand Load (k) _i(k-1) represent respectively Follow_UAV _ithe bullet-loading capacity of current time and previous moment.

$P\_{\mathrm{threat}}_{\mathrm{ij}}\left(k\right)=\underset{k=c\_t-T_s}{\overset{c\_t}{\mathrm{\&Sigma;}}}{P}_{\mathrm{dij}}\left(k\right)---\left(4\right)$

In formula, c_t is current sampling instant, T _sfor the cumulative time section of instant probability, P _dijtST _jradar to Follower_UAV _iinstantaneous tracking probability, its expression formula is suc as formula shown in (5).

$P_{\mathrm{dij}}=\left\{\begin{array}{cc}1/[1+{(c_{2j}{R_{\mathrm{ij}}}^4/{\mathrm{\&sigma;}}_i)}^{c_{1j}}]& R<R_{\max }\\ 0& R\&GreaterEqual;R_{\max }\end{array}\right.---\left(5\right)$

In formula, R _ijfollower_UAV _iwith TST _jbetween distance; c _1j, c _2jconstant, by TST _jradar type determine; σ _ifollower_UAV _iradar Cross Section (Radar Cross-Section, RCS).The RCS of UAV is the included angle X about UAV course and TST horizontal direction, the angle of pitch between TST and UAV the function of side acceleration and acceleration of gravity ratio μ, its expression formula is suc as formula shown in (6), (7), (8).

$\mathrm{\&sigma;}=\frac{\mathrm{\&pi;}\&CenterDot;a^2{b}^2{c}^2}{{(a^2{\sin }^2{\mathrm{\&lambda;}}_e{\cos }^2{\mathrm{\&mu;}}_e+b^2{\sin }^2{\mathrm{\&lambda;}}_e{\sin }^2{\mathrm{\&mu;}}_e+c^2{\cos }^2{\mathrm{\&lambda;}}_e)}^2}---\left(6\right)$

Leader_UAV is according to the cost_auction distribution of finishing the work, if Follower_UAV receives the task that Leader_UAV issues, need to calculate it and carry out the distance between TST, if this distance is less than or equal to the bee-line D_start that starts to carry out strike task, Follower_UAV starts to hit, and in original place, spirals until target is smashed; Otherwise Follower_UAV utilizes MPC algorithm to hit path planning.According to the kinematical equation of aircraft, set up nonlinear forecast model suc as formula shown in (9), (10), (11), (12), (13).

z(k+1)＝z(k)+s·sin(θ(k+1)) (11)

θ(k+1)＝θ(k)+θ ₀·u _θ(k) U _θmin≤u _θ≤U _θmax (13)

In formula, (x, y, z) represents position; S represents the air line distance between current navigation spots and next navigation spots; position angle, be constant, represent azimuthal change step, azimuthal controlled quentity controlled variable, it is the border of position angle controlled quentity controlled variable; θ is the angle of pitch, θ ₀be constant, represent the change step of the angle of pitch, u _θthe controlled quentity controlled variable of the angle of pitch, U _{θ min}, U _{θ max}it is the border of angle of pitch controlled quentity controlled variable.

u _θit is the optimum control amount obtaining by rolling optimization path cost function cost_path.The threat cost that path cost function cost_path is herein caused by state error cost, TST radar and the control cost in control domain form, and each element expression in cost_path is suc as formula shown in (14).

cost_path _ij(k)＝αc _ij ^T(k)C _ic _ij(k) (14)

+βP_threat _ij(k)+γu ^T(k)u(k)

In formula, c _ij(k) represent Follower_UAV _iwith TST _jbetween distance cost, i.e. c _ij(k)=(x _i-x _j, y _i-y _j, z _i-z _j) ^t, TST _jposition S _j(k)=(x _j, y _j, z _j) ^trepresent reference locus, α is its weights, C _ifor UAV _iand the distance cost weighting matrix between TST; P_threat _ijrepresent TST _jto Follower_UAV _ithreat ability, β is its weights; U (k) is controlled quentity controlled variable, the weights that γ is it.

Fig. 2 discloses the strike path locus figure of quick task dynamic allocation algorithm while adopting in the present invention, and by Fig. 2, we can see, UAVs forms into columns and carried out task renewal when sampling instant k=8, by Follower_UAV ₅substitute Follower_UAV ₁execution is to TST ₃strike task.Due under current task allocation result, the distance of forming into columns between interior every Follower_UAV and distribution target is all greater than D_start, so every Follower_UAV utilizes MPC algorithm to complete strike path, plans online.At k=111 moment TST ₃there is Follower_UAV ₅arrive and attack position on time, completed TST ₃strike.In like manner, Follower_UAV ₅complete TST ₃strike task after will exert an influence to the time window width of the TST also not hitting, UAVs formation task need to heavily be distributed.

The realization of quick task dynamic allocation algorithm while the invention allows for multiple no-manned plane under a kind of battlefield surroundings, comprises following steps:

Step 1:Follower_UAV submits TST observation information to Leader_UAV, and Leader_UAV produces task list, and carries out task auction according to formula (1), and the task that realizes is distributed;

Step 2:Follower_UAV is according to task allocation result, judges whether the task list of self is empty, if it is empty, and in current location waits of spiraling, until task list while be sky, is executed the task by new task list, otherwise performed step 3;

Step 3:Follower_UAV is according to the task list of self, judgement and the distance of carrying out between TST, if this is apart from being less than or equal to D_start, Follower_UAV starts to strike target, and in original place, spirals until target is smashed; Otherwise Follower_UAV carries out path planning by the optimum control amount obtaining according to formula (13), after reaching striking distance, start to strike target, and in original place, spiral until target is smashed;

Step 4:Follower_UAV carries out current task and can exert an influence to follow-up TST time window width, need to distribute and hit path and adjust UAVs formation task, will return to step 1 and carry out task weight-normality and draw, until complete the strike task of all TST.

Claims (2)

1. quick task dynamic allocation algorithm during multiple no-manned plane under battlefield surroundings, its feature is as follows, UAVs forms into columns and adopts Leader-Follower structure to carry out TST strike task, Leader_UAV mainly bears decision maker's role, Follower_UAV bears striker's role, Leader_UAV will according to time quick job change situation, and Follower_UAV executes the task the time, injure ability and hit the factors such as income and complete the distribution of formation task, on the basis of distributing in Leader_UAV task, Follower_UAV will complete flight path planning and TST will be hit according to task definition, with Leader_UAV, serve as task auctioneer, Follower_UAV serves as bidder, Follower_UAV (comprises auction price matrix cost_auction separately: and distance d between TST, carry out the time t that TST hits task, to the ability of the injuring Damage of TST and hit income VLR) send to Leader_UAV, on this basis, Leader_UAV is under the constraint of TST time window width duration, according to cost_auction, bidder is selected, the distribution of finishing the work, duration is subject to Follower_UAV to carry out the impact of current task, its expression formula is suc as formula shown in (1), each element expression in cost_auction is suc as formula shown in (2),

duration _j(k)＝duration _j(k-1)+variation _ij (1)

In formula, duration _jrepresent TST _jtime window width, variation _ijrepresent

Follower_UAV _icarry out current task to TST _jthe impact of time window width,

cost_auction _ij(k)＝d _ij(k)／max(d _ij(k))+Damage _ij(k)

+VLR _ij(k)+max(t _ij(k))/t _ij(k) (2)

+max(P_threat _ij(k))/P_threat _ij(k)

Suppose V={1,2 ..., N} is Follower_UAV quantity, in forming into columns, the quantity of UAVs is N+1, and T={1,2 ..., M} is TST quantity, in formula, and d _ijrepresent Follower_UAV _i(i ∈ V) and TST _jdistance between (j ∈ T); t _ijrepresent Follower_UAV _icarry out TST _jhit task time; Damage _ijrepresent Follower_UAV _ito TST _jthe ability of injuring; VLR _ijrepresent Follower_UAV _ihit TST _jbenefit, it is Damage _ijwith bullet-loading capacity Load _ifunction, its expression formula is suc as formula shown in (3); P_threat _ijrepresent TST _jto Follower_UAV _ithreat probabilities, will pass through TST _jradar to Follower_UAV _iinstantaneous tracking probability P _dijat cumulative time section T _son integration obtain, its expression formula is suc as formula shown in (4),

${\mathrm{VLR}}_{\mathrm{ij}}\left(k\right)={\mathrm{VLR}}_{\mathrm{ij}}(k-1)\&CenterDot;\frac{{\mathrm{Load}}_i\left(k\right)-{\mathrm{Load}}_i(k-1)}{{\mathrm{Damage}}_{\mathrm{ij}}\left(k\right)-{\mathrm{Damage}}_{\mathrm{ij}}(k-1)}---\left(3\right)$

In formula, Load _iand Load (k) _i(k-1) represent respectively Follow_UAV _ithe bullet-loading capacity of current time and previous moment,

$P\_{\mathrm{threat}}_{\mathrm{ij}}\left(k\right)=\underset{k=c\_t-T_s}{\overset{c\_t}{\mathrm{\&Sigma;}}}{P}_{\mathrm{dij}}\left(k\right)---\left(4\right)$

In formula, c_t is current sampling instant, T _sfor the cumulative time section of instant probability, P _dijtST _jradar to Follower_UAV _iinstantaneous tracking probability, its expression formula is suc as formula shown in (5),

$P_{\mathrm{dij}}=\left\{\begin{array}{cc}1/[1+{(c_{2j}{R_{\mathrm{ij}}}^4/{\mathrm{\&sigma;}}_i)}^{c_{1j}}]& R<R_{\max }\\ 0& R\&GreaterEqual;R_{\max }\end{array}\right.---\left(5\right)$

In formula, R _ijfollower_UAV _iwith TST _jbetween distance; c _1j, c _2jconstant, by TST _jradar type determine; σ _ifollower_UAV _iradar Cross Section (Radar Cross-Section, RCS), the RCS of UAV is the included angle X about UAV course and TST horizontal direction, the angle of pitch between TST and UAV the function of side acceleration and acceleration of gravity ratio μ, its expression formula is suc as formula shown in (6), (7), (8),

$\mathrm{\&sigma;}=\frac{\mathrm{\&pi;}\&CenterDot;a^2{b}^2{c}^2}{{(a^2{\sin }^2{\mathrm{\&lambda;}}_e{\cos }^2{\mathrm{\&mu;}}_e+b^2{\sin }^2{\mathrm{\&lambda;}}_e{\sin }^2{\mathrm{\&mu;}}_e+c^2{\cos }^2{\mathrm{\&lambda;}}_e)}^2}---\left(6\right)$

Leader_UAV is according to the cost_auction distribution of finishing the work, if Follower_UAV receives the task that Leader_UAV issues, need to calculate it and carry out the distance between TST, if this distance is less than or equal to the bee-line D_start that starts to carry out strike task, Follower_UAV starts to hit, and in original place, spirals until target is smashed; Otherwise Follower_UAV utilizes MPC algorithm to hit path planning, according to the kinematical equation of aircraft, set up nonlinear forecast model suc as formula shown in (9), (10), (11), (12), (13),

z(k+1)＝z(k)+s·sin(θ(k+1)) (11)

θ(k+1)＝θ(k)+θ ₀·u _θ(k) U _θmin≤u _θ≤U _θmax (13)

In formula, (x, y, z) represents position; S represents the air line distance between current navigation spots and next navigation spots; position angle, be constant, represent azimuthal change step, azimuthal controlled quentity controlled variable, it is the border of position angle controlled quentity controlled variable; θ is the angle of pitch, θ ₀be constant, represent the change step of the angle of pitch, u _θthe controlled quentity controlled variable of the angle of pitch, U _{θ min}, U _{θ max}the border of angle of pitch controlled quentity controlled variable,

u _θit is the optimum control amount obtaining by rolling optimization path cost function cost_path, the threat cost that path cost function cost_path is herein caused by state error cost, TST radar and the control cost in control domain form, each element expression in cost_path is suc as formula shown in (14)

cost_path _ij(k)＝αc _ij ^T(k)C _ic _ij(k) (14)

+βP_threat _ij(k)+γu ^T(k)u(k)

In formula, c _ij(k) represent Follower_UAV _iwith TST _jbetween distance cost, i.e. c _ij(k)=(x _i-x _j, y _i-y _j, z _i-z _j) ^t, TST _jposition S _j(k)=(x _j, y _j, z _j) ^trepresent reference locus, α is its weights, C _ifor UAV _iand the distance cost weighting matrix between TST; P_threat _ijrepresent TST _jto Follower_UAV _ithreat ability, β is its weights; U (k) is controlled quentity controlled variable, the weights that γ is it.

2. the realization of quick task dynamic allocation algorithm during multiple no-manned plane under battlefield surroundings, is characterized in that, comprises following steps:

Step 1:Follower_UAV submits TST observation information to Leader_UAV, and Leader_UAV produces task list, and carries out task auction according to formula (1), and the task that realizes is distributed;

Step 2:Follower_UAV is according to task allocation result, judges whether the task list of self is empty, if it is empty, and in current location waits of spiraling, until task list while be sky, is executed the task by new task list, otherwise performed step 3;

Step 3:Follower_UAV is according to the task list of self, judgement and the distance of carrying out between TST, if this is apart from being less than or equal to D_start, Follower_UAV starts to strike target, and in original place, spirals until target is smashed; Otherwise Follower_UAV carries out path planning by the optimum control amount obtaining according to formula (13), after reaching striking distance, start to strike target, and in original place, spiral until target is smashed;

Step 4:Follower_UAV carries out current task and can exert an influence to follow-up TST time window width, need to distribute and hit path and adjust UAVs formation task, will return to step 1 and carry out task weight-normality and draw, until complete the strike task of all TST.