CN111077900B

CN111077900B - Unmanned aerial vehicle formation control method based on task exchange

Info

Publication number: CN111077900B
Application number: CN201911199816.0A
Authority: CN
Inventors: 胡劲文; 赵春晖; 侯晓磊; 潘泉; 徐钊; 程雪梅; 刘慧霞; 祝小平
Original assignee: Northwestern Polytechnical University
Current assignee: Northwestern Polytechnical University
Priority date: 2019-11-29
Filing date: 2019-11-29
Publication date: 2021-02-09
Anticipated expiration: 2039-11-29
Also published as: CN111077900A

Abstract

The invention discloses an unmanned aerial vehicle formation control method based on task exchange, which is used for allocating a final target position and priority to each unmanned aerial vehicle in an unmanned aerial vehicle system; determining the final target position of the flight: when each unmanned aerial vehicle flies and other unmanned aerial vehicles exist in the communication range of the unmanned aerial vehicle, the optimal adjacent unmanned aerial vehicle in the communication range is selected by adopting an unmanned aerial vehicle final target position exchange method to carry out final target position exchange; executing a flight action, and repeating the steps of determining a final flight target and executing the flight action until the unmanned aerial vehicle system reaches an ideal formation; this application can realize that the unmanned aerial vehicle system converges to ideal formation fast and avoids single unmanned aerial vehicle and adjacent unmanned aerial vehicle to bump in the flight through confirming the final target position step of flight, through carrying out the flight action step, can avoid the unmanned aerial vehicle flight in-process to miss final target position, realizes that unmanned aerial vehicle system does not have the collision convergence.

Description

Unmanned aerial vehicle formation control method based on task exchange

[ technical field ] A method for producing a semiconductor device

The invention belongs to the technical field of unmanned aerial vehicle control, and particularly relates to an unmanned aerial vehicle formation control method based on task exchange.

[ background of the invention ]

In recent years, the formation control of the multi-unmanned aerial vehicle system has a wide potential application value in the fields of search and rescue, rescue tasks, logistics transportation and the like, and is gradually valued. The research subject in the field of formation control is of wide scope, including maintenance/adjustment of formation, obstacle avoidance, distributed path planning, processing of cooperation information, and the like. In a large-scale multi-drone system, one of the most basic drone formation control problems is how to achieve convergence of drone clusters to ideal formation without collision.

At present, in some existing methods, the ideal formation of unmanned aerial vehicles can be realized, the collision avoidance problem when the unmanned aerial vehicles converge to the ideal formation can also be realized, but generally, one of the unmanned aerial vehicles is selected as a director in the unmanned aerial vehicle formation, the command of the whole unmanned aerial vehicle system is carried out, the collision is avoided when the unmanned aerial vehicle system converges to the ideal formation, therefore, the director acquires the state information of all the unmanned aerial vehicles, and carries out unified planning, but, in the actual situation, the unmanned aerial vehicle communication range in the unmanned aerial vehicle system is small, the whole unmanned aerial vehicle system is difficult to cover, the control signal of the director can not be received by the individual unmanned aerial vehicle, and the ideal formation or collision avoidance is difficult to realize.

[ summary of the invention ]

The invention aims to provide a task exchange-based unmanned aerial vehicle formation control method and device, so that an unmanned aerial vehicle system can be quickly converged to an ideal formation, and collision during convergence is avoided.

The invention adopts the following technical scheme: an unmanned aerial vehicle formation control method based on task exchange comprises the following steps:

allocating a final target position and a priority to each unmanned aerial vehicle in the unmanned aerial vehicle system;

determining the final target position of the flight: when each unmanned aerial vehicle flies and other unmanned aerial vehicles exist in the communication range of the unmanned aerial vehicle, the optimal adjacent unmanned aerial vehicle in the communication range is selected by adopting an unmanned aerial vehicle final target position exchange method to carry out final target position exchange;

executing a flight action: when the unmanned aerial vehicle after the exchange of the final target positions flies to a new final target position, the flying distance of each time period is upsilon_i,k＝λ_i,k(φ_i,k-p_i,k) Wherein upsilon is_i,kIs the actual flight distance of the unmanned aerial vehicle i at the time k, and is more than or equal to 0 and less than or equal to lambda_i,k≤1，λ_i,kFor step size control factor, phi, of unmanned aerial vehicle i at time k_i,kIs the final target position, p, of drone i at time k_i,kIs the current position of the unmanned aerial vehicle i at time k;

and repeating the steps of determining the final target of the flight and executing the flight action until the unmanned aerial vehicle system reaches the ideal formation.

Further, the step size control factor

Wherein Γ (a, b) (a > 0, b > 0) is a saturation function,

Z_i,k＝μU，Z_i,kthe step length of the movement of the unmanned aerial vehicle i in the single control cycle is U, the maximum flight distance of the unmanned aerial vehicle i in the single control cycle is more than 0 mu and less than or equal to 1, and mu is a step length scale factor.

Further, determining the flight final target position specifically includes:

acquiring the current position, the final target position and the priority of the current unmanned aerial vehicle;

sending broadcast information to the unmanned aerial vehicle group, and receiving broadcast information sent by other unmanned aerial vehicles in the unmanned aerial vehicle group; the broadcast information comprises a current position, a final target position and a priority;

when the current unmanned aerial vehicle meets a first preset condition, generating an unmanned aerial vehicle blocking set; wherein the first preset condition is as follows: the distance between the final target position and the current position is larger than the radius of the attraction domain of the final target position and may prevent the unmanned aerial vehicle set from being an empty set,

selecting an optimal adjacent unmanned aerial vehicle according to the set of the blocking unmanned aerial vehicles, and sending request information for exchanging the final target position to the optimal adjacent unmanned aerial vehicle;

and after receiving the information of the optimal adjacent unmanned aerial vehicle for agreeing to exchange, taking the final target position of the optimal adjacent unmanned aerial vehicle as the final target position of the current unmanned aerial vehicle.

Further, hinder unmanned aerial vehicle set to pass through B_i,k＝{j∈N_i,k|φ_i,k∈S_j,i,k,||p_i,k-p_j,k||≤R_SWITCHGeneration;

wherein, B_i,kSet phi of blocking unmanned aerial vehicles for current unmanned aerial vehicle i at current moment k_i,kIs the final target position of the current unmanned aerial vehicle i at the current moment k, S_j,i,kVoronoi region for drone j, R_SWITCHIs a first predetermined distance value.

Further, the method for selecting the optimal adjacent unmanned aerial vehicle according to the set of blocking unmanned aerial vehicles comprises the following steps:

respectively generating a first blocking unmanned aerial vehicle subset and a second blocking unmanned aerial vehicle subset according to the blocking unmanned aerial vehicle set; the first subset of hindering drones is:

wherein the content of the first and second substances,

for the first subset of hindering drones for the current drone i at the current moment k, A_j,kIs the final target position phi of drone j_j,kThe attraction domain of (1);

the second subset of hindering drones is:

wherein the content of the first and second substances,

a second subset of hindering drones for the current drone i at the current time k;

when in use

When passing through

Selecting an optimal adjacent unmanned aerial vehicle; wherein j is the optimal adjacent unmanned aerial vehicle;

when in use

Then, it is determined whether j ∈ B exists_i,kSatisfies p_i,k∈A_i,k；

If present, by G_i,k＝{g∈N_i,k|||p_g,k-p_j.k||≤R_STOP,p_g,k∈A_j,k,θ_g,k＞θ_i,kGenerate a third set of drones, and pass

Selecting an optimal adjacent unmanned aerial vehicle;

wherein g is another neighboring drone, p, of the current drone i_g,kIs the position of the unmanned plane g at the current moment k, theta_g,kThe priority of the unmanned plane g at the current time k.

Further, it may be hindered from concentrating into:

D_i,k＝{j∈N_i,k|θ_j,k＜θ_i,k,||p_i,k-p_j,k||≤2R_STOP}，

wherein D is_i,kFor a set of possible blocking drones of current drone i at current moment k, j is the neighboring drone of current drone i, N_i,kSet of neighboring drones, theta, of drone i at current moment k_j,kIs the priority of drone j, θ_i,kPriority of current drone i, p_i,kIs the position of the unmanned aerial vehicle i at the current moment k, p_j,_kIs the position of the unmanned plane j at the current moment k, R_STOPIs the collision threshold.

The invention has the beneficial effects that: according to the method and the device for determining the final target position of the unmanned aerial vehicle, the unmanned aerial vehicle system can be rapidly converged to an ideal formation by determining the step of the final target position of flight, collision between a single unmanned aerial vehicle and an adjacent unmanned aerial vehicle in flight is avoided, the flight strategy of each time period of the unmanned aerial vehicle can be optimized by executing the step of flight action, the final target position is prevented from being missed in the flight process of the unmanned aerial vehicle, a plurality of buffer areas are formed, and the unmanned aerial vehicle can be sequentially connected to realize smoother change of speed.

[ description of the drawings ]

Fig. 1 is a flowchart of a task exchange-based unmanned aerial vehicle formation control method according to the present application;

FIG. 2 is a flow chart of determining a final target position for a flight in an embodiment of the present application;

FIG. 3 is another flow chart of the present embodiment of determining a final target position for flight;

FIG. 4 is a schematic diagram of the voronoi region division in the embodiment of the present application;

fig. 5 is a schematic view of the starting positions of five unmanned aerial vehicles in the embodiment of the application;

fig. 6 is a schematic diagram of end positions and path trajectories of five unmanned aerial vehicles in an embodiment of the application.

[ detailed description ] embodiments

The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

The method of the embodiment of the application is applied to each unmanned aerial vehicle in the unmanned aerial vehicle system, and can be used for each aircraft in other aircraft systems. When the unmanned aerial vehicle system carries out ideal formation convergence, because every unmanned aerial vehicle's final target location is different, the current position is different, if not manage unmanned aerial vehicle, when two unmanned aerial vehicle's final target location and current position intercrossing, these two unmanned aerial vehicle are when the motion of orientation respective final target location, it collides possibly on the way in the motion, this is very big damage to unmanned aerial vehicle, not only be difficult to realize the ideal formation, lead to unmanned aerial vehicle to damage more likely, and then lead to whole unmanned aerial vehicle system paralyse. Therefore, when the unmanned aerial vehicle system performs formation convergence, each unmanned aerial vehicle is controlled, and the motion trail or the path point of each unmanned aerial vehicle is planned.

An embodiment of the present application provides a task exchange-based unmanned aerial vehicle formation control method, as shown in fig. 1, specifically including the following steps:

and step S10, allocating a final target position and priority to each unmanned aerial vehicle in the unmanned aerial vehicle system.

Step S20, determining the final flying target position: when each unmanned aerial vehicle flies and other unmanned aerial vehicles exist in the communication range of the unmanned aerial vehicle, the optimal adjacent unmanned aerial vehicle in the communication range is selected by adopting an unmanned aerial vehicle final target position exchange method to carry out final target position exchange.

Step S30, performing a flight action: performing final target position intersectionWhen the changed unmanned aerial vehicle flies to a new final target position, the flying distance of each time period is upsilon_i,k＝λ_i,k(φ_i,k-p_i,k) Wherein upsilon is_i,kIs the actual flight distance of the unmanned aerial vehicle i at the time k, and is more than or equal to 0 and less than or equal to lambda_i,k≤1，λ_i,kIs the distance weight factor, phi, of unmanned aerial vehicle i at time k_i,kIs the final target position, p, of drone i at time k_i,kIs the current position of drone i at time k.

And step S40, repeating the steps of determining the final flight target and executing the flight action until the unmanned aerial vehicle system reaches the ideal formation. When the unmanned aerial vehicle system does not reach the ideal formation, the operation returns to the step S20 to continue the execution.

According to the method and the device for determining the final target position of the unmanned aerial vehicle, the collision between the single unmanned aerial vehicle and the adjacent unmanned aerial vehicle in the flight process can be avoided through the step of determining the final target position of the flight, the flight strategy of each time period of the unmanned aerial vehicle can be optimized through the step of executing the flight action, the final target position of the unmanned aerial vehicle in the flight process is avoided from being missed, a plurality of buffer areas are formed, and the unmanned aerial vehicle can be connected in sequence to realize the smoother change of the speed.

In the process of determining the final target position of the flight, the formation control problem is converted into a general real-time optimization problem, wherein N final target positions are respectively allocated to N drones, and the movement route or path point of each drone needs to be calculated, which is a typical NP problem.

Moreover, a distributed control method is provided in the embodiment of the application, efficient real-time calculation of suboptimal solutions is realized by introducing the priority of the unmanned aerial vehicle, which is similar to a general idea of solving an NP problem by using a sequential implementation method. However, the embodiment of the application is realized by parallel or distributed operation of each unmanned aerial vehicle. Wherein each drone decides whether to exchange the final target position with an adjacent drone and calculates where to move in real time according to local information.

By integrating the control method and the strategy, the method for determining the final flight target position of the embodiment is provided, and the flow of the method is shown in fig. 2 and fig. 3, wherein the method is an unmanned aerial vehicle system based on a task exchange distributed control algorithm. Finally, a detailed parameter design process is provided in this embodiment, which is to facilitate the user to design and control the minimum separation distance, the moving speed, the communication range, the initial setting, the ideal formation, and so on.

Through the method of this embodiment, under the condition that satisfies minimum separation distance, unmanned aerial vehicle's initial position can set up at will, and is unrestricted, and can prove that ideal formation can realize gradually. Finally, the effectiveness of the algorithm under the random initial value is verified through Monte Carlo simulation, and the result shows that the method has good collision prevention performance and can quickly generate the target at exponential speed.

Before describing the specific implementation steps, some basic definitions are first given. N drones of the same type are defined as the set V ═ 1, 2.., N }, which have the same communication and mobility capabilities, such as the same communication range R_CAnd a maximum moving speed U. The dynamic waypoint model for each drone is defined as:

p_i,k+1＝p_i,k+υ_i,k，

where i denotes the drone, k denotes the discrete time index, p_i,k+1For the position of unmanned plane i at time k +1, p_i,kIs the position of unmanned plane i at time k, upsilon_i,kDisplacement generated by the movement of the unmanned aerial vehicle i from the moment k to the moment k +1, | | upsilon_i,kAnd | | is less than or equal to U, and U is the maximum moving step length of the unmanned aerial vehicle in the unit time period. If and only if p_i,k-p_j,k||≤R_CAnd the unmanned aerial vehicle i and the unmanned aerial vehicle j can realize mutual communication at the moment k. Thus, the topology of the drone system is represented as: ζ ═ ε, upsilon, ε_k＝{{i,j}|||p_i,k-p_j,k||≤R_C,j≠i,(i,j)∈V}。

Defining a set of drones adjacent to drone i as N_i,k＝{j|{i,j}∈ε_kJ ∈ V }, the ideal formation of the drone system is defined as the set of final target locations Q ═ Q ∈ V }, where the set of the final target locations Q is defined as the set of the target locations Q₁,q₂,…,q_NWherein q is_m≠q_nM is not equal to n, and m belongs to V. Converge to ideal formation if every final target position in Q is occupied by every drone in V. This can be expressed as a mapping φ: v → Q, indicates that each drone has a final target position. Phi is a_i,_ke.Q represents the final target position allocated by the unmanned aerial vehicle i at the moment k, and represents the task allocation phi of a group of final target positions_k＝[φ_1,k,φ_2,k,…,φ_N,k]. For position vectors

For indicating, displacement vectors

And (4) showing. Thus, a set of allowed drone displacements is defined as

Based on the above definition, the formation control problem is transformed into the following mathematical optimization problem:

wherein the content of the first and second substances,

the embodiment of determining the final target position of the flight according to the embodiment of the present application, as shown in fig. 2, specifically includes the following steps:

step S110, obtaining the current position p of the current unmanned aerial vehicle_i,kFinal target position phi_i,kAnd priority theta_i,k. Here, the current position may be acquired by GPS navigation or the like, or may be initially input by the user. And when the final target position is reached, the user acquires the final target position of each unmanned aerial vehicle according to the formation of the unmanned aerial vehicle system. The priority is specified by the user when specifying the final target position for each unmanned aerial vehicle, and the current time is judged according to the priority for the convenience of later stageThe drone is bound to the final target position with which the neighboring drone exchanges the final target position set, that is, when the final target position of the drone changes, the priority of the drone changes accordingly.

There are three basic issues to consider for task information exchange of drones:

1) is an exchange required? 2) Who is the exchange information sent? 3) Which should be agreed upon if multiple information exchange requests are received? The biggest difficulty in solving the above problem is that these decisions have to be made separately due to the limited communication range. This may result in different decisions canceling each other out rather than producing a common optimal decision in a centralized optimization.

To solve the above problem, the present embodiment introduces a target position phi corresponding to each final target position_i,kRelative priority, also understood as priority of the drone, θ_i,kE.g., V ≠ j, θ for i ≠ j ≠ N }, e.g., 1,2_i,k≠θ_j,k. Thus, θ is a mapping of time k, V, to V, the former V representing the set of drones and the latter V representing the set of final target location priorities. Theta_i,kThe smaller, the destination phi_i,kThe smaller the priority.

Step S120, sending broadcast information to the unmanned aerial vehicle group, and receiving broadcast information sent by other unmanned aerial vehicles in the unmanned aerial vehicle group; wherein the broadcast information includes a current location, a final target location, and a priority. Broadcast information of sending and receiving is all carried out at the communication distance scope, only carry on local communication can, need not carry out the communication of whole unmanned aerial vehicle system to reduce the requirement to the hardware, reduce whole calculated amount.

Step S130, when the current unmanned aerial vehicle meets a first preset condition, generating a set of blocking unmanned aerial vehicles; wherein the first preset condition is as follows: the distance between the final target position and the current position is larger than the radius of the attraction domain of the final target position, and the unmanned aerial vehicle set can be prevented from being an empty set.

In this embodiment, in the neighboring drones that may hinder the set of drones from being the current drone, there is a set of drones that may hinder the next movement of the drone, and the drones in this set first satisfy that they are drones neighboring to drone i, and secondly, need to satisfy the priority that their priority is smaller than drone i, and the distance between them and drone i is less than or equal to twice the collision critical value, which is specifically expressed as:

D_i,k＝{j∈N_i,k|θ_j,k＜θ_i,k,||p_i,k-p_j,k||≤2R_STOP}，

wherein D is_i,kFor a set of possible hindering drones for current drone i at current time k, j is the neighboring drone of current drone i, θ_j,kIs the priority of drone j, θ_i,kPriority of current drone i, p_i,kIs the position of the unmanned aerial vehicle i at the current moment k, p_j,kIs the position of the unmanned plane j at the current moment k, R_STOPIs the collision threshold.

When the first preset condition is met, the distance from the current unmanned aerial vehicle to the final target position is far, and in the process of moving to the final target position, collision with some adjacent unmanned aerial vehicles may occur. Under this condition, it is necessary to calculate whether there is another method to avoid the occurrence of the collision. Therefore, firstly, a set of blocking drones is established through a certain rule, and all drones in the set collide with the current drone when the drone moves.

Hinder unmanned aerial vehicle set in this embodiment to pass through B_i,k＝{j∈N_i,k|φ_i,k∈S_j,i,k,||p_i,k-p_j,k||≤R_SWITCHIs generated, wherein N_i,kNeighboring set of unmanned aerial vehicles, B, of unmanned aerial vehicle i at current moment k_i,kSet phi of blocking unmanned aerial vehicles for current unmanned aerial vehicle i at current moment k_i,kIs the final target position of the current unmanned aerial vehicle i at the current moment k, S_j,i,kVoronoi region for drone j, R_SWITCHIs a first predetermined distance value.

At this obstacle unmanned aerial vehicle concentration, every unmanned aerial vehicle must be unmanned aerial vehicle i's adjacent unmanned aerial vehicle to guarantee that current unmanned aerial vehicle only communicates with adjacent unmanned aerial vehicle, reduce the calculated amount. Moreover, the unmanned aerial vehicle in the set is located in the voronoi area, namely the current unmanned aerial vehicle, so that the movement distance of the unmanned aerial vehicle can be correspondingly reduced. Finally, the distance between the unmanned aerial vehicles in the concentration of the unmanned aerial vehicles and the current unmanned aerial vehicle is smaller than a preset threshold value of the final destination position of the exchange of the two unmanned aerial vehicles, namely a first preset distance value.

For the division of the Vono area of each unmanned aerial vehicle, the specific method comprises the following steps:

first, define the whole movable space as a set

This set contains all the final target positions in Q and phi₀All initial final destination location assignment schemes. For i, j ∈ V, C_i,_j,k＝{s∈|||s-p_i,k||≤||s-p_j,kIf this is the case, then this formula is further defined as

i and j are the serial numbers of the unmanned aerial vehicles respectively,

wherein, as shown in FIG. 3, C_i,j,k∩C_j,i,kIs represented by C_i,j,kAnd C_j,i,kA common boundary. Satisfy | | | s-p simultaneously_i,k||＝||s-p_j,k||，s∈C_i,j,k∩C_j,i,k. Is defined so that

And

this implies that

And

{S_1,k,S_2,k,...,S_N,kis (d) covered withIs called as

Veno partitioning of the Domain, S_i,kIs said to contain position p_i,kThe voronoi division of drone i.

Note that S_i,kIs calculated by relying on all S_i,j,kAnd j ∈ V. All S_i,j,kAre difficult to obtain in distributed networks with limited communication distances. From the perspective of each drone i, each local voronoi zoning needs only local information defined as follows:

it shows for j ∈ N_i,k，

That is to say for some cases where j ≠ i,

and

with an overlapping portion therebetween. However, if i, j two drones collide, that is to say for some δ < R_C，||p_i,k-p_j,kAnd | l < delta. They should be adjacent to each other, which means that

Therefore, the temperature of the molten metal is controlled,

can regard as unmanned aerial vehicle i's work area, this regional calculation only needs local information. Note that each unmanned aerial vehicle i can only calculate the action area of the unmanned aerial vehicle i

Can also calculate when j is equal to N_i,kS of (1)_i,j,k。

Step S140, after the set of blocking drones is obtained, one drone needs to be selected and replaced with the final target position, so as to avoid collision between two drones. Therefore, according to the set of hindering drones, the optimal adjacent drone is selected according to a predetermined decision or condition and sent with the request information to exchange the final target position. By selecting the optimal neighboring drones, further calculations can be avoided and the probability of collisions will be minimized, allowing the drone system to converge to an ideal formation without collisions.

In this embodiment, the method for selecting the optimal adjacent drone according to the set of blocking drones is as follows:

a first subset of hindering drones and a second subset of hindering drones are generated from the set of hindering drones, respectively.

The first subset of hindering drones is:

wherein the content of the first and second substances,

for the first subset of hindering drones for the current drone i at the current moment k, A_j,kIs the final target position phi of drone j_j,kAttraction domain of A_i,k＝{s∈|||s-φ_i,k≤R_ATTR||}。

The second subset of hindering drones is:

wherein the content of the first and second substances,

when in use

When passing through

Selecting an optimal adjacent unmanned aerial vehicle; wherein j is the optimal adjacent drone.

The above distance may be defined as follows: r_C≥R_SWITCH≥2R_STOP，R_ATTR≥2R_STOPAt the same time

Since Q does not change with time, for all i ∈ V, φ_i,k∈Q，

Indicating that for all i e V,

to satisfy this, make

d is the distance between adjacent drones in the ideal formation. Approximately, define the minimum distance between actual drones as

Each unmanned plane j belongs to B_i,kIs a final target position phi away from the unmanned aerial vehicle i_i,kMore recently, blocking drones.

Consisting of blocking drones whose destination priority is low and which are outside their desired attraction domain.

By being within their attraction domain and by p thereof_i,kObstacle drones also within the attraction domain. However, it is not limited to

Show that

That is to say, a drone can only be in one attraction area at most.

When in use

Then, it is determined whether j ∈ B exists_i,kSatisfies p_i,k∈A_i,k；

Selecting an optimal adjacent unmanned aerial vehicle;

Step S150, after sending the request information to the optimal neighboring drone, the optimal drone also selects whether to perform final target position replacement with the current drone according to a predetermined policy of the optimal drone. If the most recent drone, through calculation, has a final target location change with the current drone, with a reduced probability of collision, or no collision, the request will be received and an agreement to exchange information will be sent.

The current unmanned aerial vehicle in this embodiment receives the reply of the consent exchange message from the optimal neighboring unmanned aerial vehicle, and after receiving the reply, takes the final target position of the optimal neighboring unmanned aerial vehicle as the final target position of the current unmanned aerial vehicle, and performs flight planning according to the final target position.

According to the unmanned aerial vehicle formation method and device, the set of blocking unmanned aerial vehicles is constructed, the optimal adjacent unmanned aerial vehicles are selected according to the set of blocking unmanned aerial vehicles, the optimal adjacent unmanned aerial vehicles exchange the final target position, the final target position continues to move to an ideal formation, and collision between the unmanned aerial vehicles and the adjacent unmanned aerial vehicles in the process of moving to the ideal formation is avoided. Meanwhile, the optimization through the method can reduce subsequent calculation amount and reduce the hardware requirement of the unmanned aerial vehicle.

In another way of determining the final target position in the present embodiment, as shown in fig. 3, steps S210 to S250 in this embodiment are the same as steps S110 to S150 in the previous embodiment.

And step S241, when the current unmanned aerial vehicle does not meet the first preset condition and receives request information for exchanging the final target sent by other unmanned aerial vehicles, sending information for agreeing to exchange to other unmanned aerial vehicles, and taking the final target positions of other unmanned aerial vehicles as the final target position of the current unmanned aerial vehicle.

In this case, two cases are involved, one being that only a request message from a neighboring drone is received, and at this point the consent exchange message can be sent directly to it.

In another case, when exchange final destination request information issued by a plurality of other drones is received, the drone with the smallest priority is selected as the optimal adjacent drone. And sending the agreement exchange information to the optimal adjacent unmanned aerial vehicle.

Step S251, after sending the information of agreement to exchange, taking the final target position of the optimal neighboring drone (if the first case is, the drone sending the request information is considered to be the optimal neighboring drone) as the final target position of the current drone.

Step S252, in this embodiment, if the first preset condition is not satisfied, and the request information sent by other neighboring unmanned aerial vehicles is not received, it is determined that the current unmanned aerial vehicle does not collide with other neighboring unmanned aerial vehicles when moving to the current final target position, and at this time, the original final target position is maintained.

For the above two embodiments, after taking the final target position of the optimal adjacent drone as the final target position of the current drone, the method further includes:

taking the priority of the optimal adjacent unmanned aerial vehicle as the priority of the current unmanned aerial vehicle;

and taking the new final target position, the priority and the current position as new broadcast information, and sending the new broadcast information to the unmanned aerial vehicle group.

Through this step, can guarantee except that other adjacent unmanned aerial vehicles in current adjacent unmanned aerial vehicle set receive the information that current unmanned aerial vehicle and the optimum continuous unmanned aerial vehicle exchanged final target location to make and calculate the formula at the next moment, provide up-to-date unmanned aerial vehicle state information for other adjacent unmanned aerial vehicles, avoid taking place the condition of calculating the mistake.

It should be noted that in both embodiments of the present application it is shown that if one drone decides to send an exchange request, at the same time it will not accept any exchange requests from neighbouring drones. Conversely, if a switch request is received from a neighboring drone, it will not send a switch request. The design is to avoid the following situations: at the same time, drone i sends an exchange request to drone j, which in turn sends an exchange request to drone l. Whether drone j exchanges information with drone i depends on the decision of drone l, and therefore their decisions cannot be performed in parallel and decentralised.

In practical application, if a final target position is determined by one unmanned aerial vehicle, the unmanned aerial vehicle needs to move towards the final target position, and when j ∈ N of the unmanned aerial vehicle exists_i,kSatisfies phi_i,k∈S_j,i,kAnd p_i,k-p_j,k||≤R_STOPLet u_i,k＝0。

This is in order to guarantee that when two unmanned aerial vehicles are very close to each other, as long as one unmanned aerial vehicle probably gets into the mission area of adjacent unmanned aerial vehicle, that is to say probably collides with adjacent unmanned aerial vehicle, and this unmanned aerial vehicle will stop moving. In other words, the drone is obstructed by an adjacent drone like an obstacle. At this time, the drone i needs to find a path to bypass the neighboring drone. However, finding the optimal path is a NP problem for all drones, since the movements of each drone are mutually affected. In fact, the reason for this obstacle is that the final target location assignment determines where they should go.

By means of the two embodiments described above, their movements can be re-planned so that they can avoid such obstacles. In the step of performing the flight action in the embodiment of the present application, when the movement is allowed, it is also necessary to design a suitable motion control strategy. Given a destination phi_i,kThe most direct behavior control of drone i at time k is along a straight line phi_i,kAnd (6) moving. I.e. v is_i,k＝λ_i,k(φ_i,k-p_i,k)，0≤λ_i,k≤1，λ_i,kStep length control factor of unmanned aerial vehicle i at time k, if collision does not occur, lambda is controlled_i,k＝Γ(||φ_i,k-p_i,k||,Z_i,k) Wherein Z is_i,kFor the step size of the movement of a single control cycle of drone i, Γ (a, b) (a > 0, b > 0) is defined as the saturation function as follows:

according to the content, the embodiment of the application also relates to an unmanned aerial vehicle motion control strategy, which specifically comprises the following steps:

when two unmanned aerial vehicles position is close, satisfy the following condition promptly: existence of unmanned plane j epsilon F_i,kMu is more than 0 and less than or equal to 1 and F_i,k＝{j∈N_i,k|R_STOP＜||p_i,k-p_j,k||≤R_STOP+2U,(φ_i,k-p_i,k)^T(p_j,k-p_i,k) > 0}, let Z be_i,kμ U, wherein Z_i,_kThe step size of the movement of the unmanned aerial vehicle in a single control cycle is represented, and mu represents a step size scale factor and is used for adjusting and slowing down the movement speed of the unmanned aerial vehicle. Otherwise, let Z_i,k＝U。

Mu < 1 is set to reduce the moving speed of the drone when the drone satisfies the above conditions. R_STOP＜||p_i,k-p_j,k||＜R_STOP+2U may be regarded as oneA buffer area where the drone needs to decelerate. In fact, when the designed μ is too small, multiple buffer regions may be connected in sequence to achieve a smoother change in speed. When H buffer regions are provided, H may be set to R1, 2_STOP+2μ_h-1U＜||p_i,k-p_j,k||≤R_STOP+2μ_hAnd U is adopted. The scale factor mu associated therewith_hShould satisfy 0 ═ mu₀＜μ_h＜μ_h+1Less than or equal to 1. If it is not

Is located in the h buffer area, and wherein

Set up Z_i,k＝μ_hU。

In addition, in the above-described motion control, the ideal formation can be achieved without collision. The expression by formula is that for all k > 0, i, j belongs to V, i is not equal to j, when | | p_i,k-p_j,kWhen the ratio of the ratio is more than 0,

comparing the method in the above embodiment with two typical methods, 1) all drones move to the destination all in a straight line, regardless of task exchange and collisions; 2) regardless of task exchanges and collisions, drones move to destinations in a one-by-one manner. Method 1 has the shortest convergence time

Is taken as the lower bound. Method 2 has the longest convergence time

Is taken as the upper bound. The actual convergence time lies between the two.

When | | | p_i,k-p_j,kWhen | is greater than 0, collision does not occur. This is not secure enough for practical applications, as unmanned aerial vehicles do not haveThe recipe is treated as a particle. That is, in view of safety issues, it is necessary to ensure that for any desired separation distance δ > 0, | | p_i,k-p_j,k| > δ. Therefore, all parameters should be designed based on this.

The general idea is to design a suitable threshold value R_STOPIs greater than delta, makes p | | |_i,k-p_j,k||≤R_STOPAlways indicates p_i,k+1-p_j,k+1||≥||p_i,k-p_j,kL. At the same time, it is necessary to design an action control strategy, for example | | | p_i,k-p_j,k||≥R_STOPAlways indicates p_i,k+1-p_j,k+1| > δ. Then, d can be obtained for all k > 0, i, j ∈ V, i ≠ j₀Is greater than delta to indicate p_i,k-p_j,k| > δ. Two detailed parameter design programs are shown in tables 1 and 2, respectively. The user can select one according to the requirement. By simple calculation, it can be found that any condition that is satisfied in one program is also satisfied in another program. So that only one of the programs needs to be considered for the analysis. Before designing a new control rule, the following set is defined:

E_i,k＝{j∈N_i,k|||p_j,k-p_i,k||≤R_STOP}。

set upsilon_i,kWhen is equal to 0

And there is no man-machine j ∈ E_i,kAny one of the following conditions is satisfied:

1)φ_j,k∈S_j.i,kand | | phi_i,k-p_i,k||≤R_ATTR；

2)φ_j,k∈S_j,i,k，θ_i,k＞θ_j,kAnd | | phi_j,k-p_j,k||＞R_ATTR；

3)φ_j,k∈S_j,i,k，(φ_i,k-p_i,k)^T(p_j,k-p_i,k) Greater than 0 and | phi_i,k-p_i,k||≤R_ATTR-R_STOP；

V in the method of the present embodiment is set by adding the above-mentioned conditions_i,kWith 0 and the parameters in table 1, the following conclusions can be reached:

for all k > 0, i, j belongs to V, i is not equal to j, when | | | p_i,k-p_j,kIdeal formation can be gradually achieved when | is > δ.

Table 1 parameter design procedure 1

Table 2 parameter design program 2

As can be seen from the above description of the embodiments, the embodiments of the present application eliminate errors and very strict preconditions, and enable clusters of unmanned aerial vehicles of the same type to achieve ideal formation in a distributed manner without collision under the condition of limited communication distance and movement capability. The method only needs to impose a minimum distance limit on the initial position of the unmanned aerial vehicle cluster, and most importantly, provides detailed parameter design so that a user can accurately control the minimum separation distance between any two unmanned aerial vehicles, and meanwhile, can ensure that ideal formation can be realized.

The application example is as follows:

in the embodiment, 20 experiments are carried out in a region with the size of 15m × 15m, and the initial position of the unmanned aerial vehicle is selected randomly in each experiment. In the experiment, 5 Intel Aero RTF unmanned aerial vehicles are used as a test platform, and are set to be positioned on the same horizontal plane with the ground height of 5m, and the flight speed of the unmanned aerial vehicles does not exceed 0.5m/s at most. The control frequency for each aircraft control system is 25Hz and the waypoints at the next time are refreshed at intervals of every 0.04 s. The position of each drone is obtained by a UWB wireless location system.

In outdoor environments, UWB positioning systems can obtain more stable and accurate position results than GPS systems. The frequency of the UWB system for updating the position information of the unmanned aerial vehicles is 50Hz, and the maximum position error between the unmanned aerial vehicles does not exceed 10cm under the condition that the limiting condition that the distance between each receiver and each anchor is less than 130m is met. In this experiment, the UWB anchors were mounted on a tripod at the four vertex regions of the rectangle.

As shown in fig. 5, the diagram shows the starting positions of five drones, and the five drones are marked by triangles, as shown in fig. 6, the diagram shows the ending positions and the path trajectories of the five drones. See table 3 for experimental parameters. In this embodiment, they would be directed to different end positions, respectively, as indicated by the circles. In order to better embody the superiority of the method proposed in the embodiment, the initial position of the drone should be selected to satisfy the following conditions: if the task end point of each unmanned aerial vehicle is not switched, the unmanned aerial vehicles collide. In the above-described conditions where collision cannot be avoided, 10 experimental results were recorded in this example.

After a part of the drones have performed the task switching target, five drones finally reach the area marked by the circle (the broken line represents the motion trajectory of the drone).

The experimental results are as follows:

the movement tracks of five unmanned aerial vehicles are recorded by using one unmanned aerial vehicle at the height of 20m through a downward-looking camera, and the initial positions of the unmanned aerial vehicles in each experiment are not identical. From fig. 6 it can also be seen that the images of the trajectories of the drones (marked by broken lines) at different times are correlated with their target points.

TABLE 3 parameter set-up for the experiment

Priority and each target point phi of five unmanned aerial vehicles_i,k：θ_i,k∈V＝{1,2,3,4,5}。

Many task exchanges may occur per experiment, for example, when the distance between drones 2 and 3 is less than R_switchThen a task exchange will take place.

According to the method of the embodiment of the application, when two unmanned aerial vehicles meet the condition of task exchange, one unmanned aerial vehicle sends an exchange request to the other unmanned aerial vehicle, and the two unmanned aerial vehicles exchange task end points and finally reach a target place. As can be seen from fig. 6, when a collision is about to occur, two drones perform task exchange, and finally five drones all fly to a desired convergence position.

Claims

1. An unmanned aerial vehicle formation control method based on task exchange is characterized by comprising the following steps:

the step size control factor lambda_i,k＝Γ(||φ_i,k-p_i,k||,Z_i,k) Wherein, gamma is(a, b) (a > 0, b > 0) is a saturation function,

Z_i,k＝μU，Z_i,kthe step length of the movement of the unmanned aerial vehicle i in the single control cycle is U, the maximum flight distance of the unmanned aerial vehicle i in the single control cycle is more than 0 mu and less than or equal to 1, and mu is a step length scale factor;

2. The method for controlling formation of unmanned aerial vehicles based on task switching according to claim 1, wherein determining the final target position of flight specifically comprises:

sending broadcast information to the unmanned aerial vehicle group, and receiving broadcast information sent by other unmanned aerial vehicles in the unmanned aerial vehicle group; wherein the broadcast information comprises a current location, a final target location and a priority;

when the current unmanned aerial vehicle meets a first preset condition, generating an unmanned aerial vehicle blocking set; wherein the first preset condition is as follows: the distance of the final target position from the current position is greater than the radius of the attraction domain of the final target position and may prevent the drone set from being an empty set,

3. The method for controlling formation of unmanned aerial vehicles based on task exchange as claimed in claim 2, wherein said set of blocking unmanned aerial vehicles passes through B_i,k＝{j∈N_i,k|φ_i,k∈S_j,i,k,||p_i,k-p_j,k||≤R_SWITCHGeneration;

4. The method for controlling formation of unmanned aerial vehicles based on task switching according to claim 3, wherein the method for selecting the optimal adjacent unmanned aerial vehicle according to the set of blocking unmanned aerial vehicles comprises:

wherein the content of the first and second substances,

the second subset of hindering drones is:

wherein the content of the first and second substances,

when in use

When passing through

Selecting an optimal adjacent unmanned aerial vehicle; wherein j is^*Is the optimal adjacent unmanned plane;

when in use

Then, it is determined whether j ∈ B exists_i,kSatisfies p_i,k∈A_i,k；

Selecting an optimal adjacent unmanned aerial vehicle;

5. The method for controlling formation of unmanned aerial vehicles based on task exchange according to claim 4, wherein the set of possible blocking unmanned aerial vehicles is:

D_i,k＝{j∈N_i,k|θ_j,k＜θ_i,k,||p_i,k-p_j,k||≤2R_STOP}，

wherein D is_i,kFor a set of possible blocking drones of current drone i at current moment k, j is the neighboring drone of current drone i, N_i,kSet of neighboring drones, theta, of drone i at current moment k_j,kIs the priority of drone j, θ_i,kPriority of current drone i, p_i,kIs the position of the unmanned aerial vehicle i at the current moment k, p_j,kIs the position of the unmanned plane j at the current moment k, R_STOPIs the collision threshold.