CN110488866B - Unmanned aerial vehicle formation obstacle avoidance method based on gradient function - Google Patents

Abstract

The invention relates to an obstacle avoidance method for UAV formation based on a gradient function. (1) prepare several UAVs; (2) specify a starting position and a target position, and there are no obstacles around the target position; , each aircraft moves to its respective target position, and every time a certain aircraft moves one step, scan the global aircraft position; (4) store the value obtained by the latest scan in step (3), and prepare for the next update; (5) according to the latest storage data, calculate the distance between two adjacent aircraft, and judge whether any aircraft meet the obstacle avoidance conditions, that is, whether the closest distance is close to the safe distance; (6) enter the formation flight state, assuming that the obstacle avoidance conditions are met, the obstacle avoidance process is established. model; repeat steps (3), (4), and (5), and adopt step (6) for the obstacle avoidance strategy. The invention better realizes the dynamic obstacle avoidance within the aircraft group and the aircraft to the aircraft, and makes up for the deficiencies in the traditional dynamic obstacle avoidance algorithm.

Description

A UAV formation obstacle avoidance method based on gradient function

technical field

The invention relates to the field of UAV formation obstacle avoidance and formation control, in particular to an UAV formation obstacle avoidance method based on a gradient function.

Background technique

With the continuous development of current aerospace technology, the application of UAVs in various fields of society is becoming more and more extensive. In the occasions where the demand is constantly being excavated, the application of a single UAV is more and more limited. For example, when performing a certain task, the labor cost and time cost of a single drone are much greater than that of multiple drones. As a result, the task scope of a single aircraft is narrow, resulting in problems such as low execution efficiency.

Obstacle avoidance is a popular problem in the field of multi-UAV formations. Most of the traditional methods are based on two-dimensional planes, which are improvements to existing algorithms. There is no reliable model for such problems or problems in the model. Summarize.

SUMMARY OF THE INVENTION

The invention proposes a tentative model for the obstacle avoidance problem between the aircraft in the aircraft group in obstacle avoidance. Model and back-collision (rear-end) model. The proper explanation is given for the model, which solves the obstacle avoidance problem in a multi-UAV swarm in a two-dimensional plane, and can be simulated and applied through experiments.

The present invention is achieved through the following technical solutions:

A UAV formation obstacle avoidance method based on gradient function, including the following steps

(1) Prepare several drones;

(2) Specify the starting position and the target position;

a. Determine the starting position of the drone fleet;

b. Determine the target location of the drone fleet;

c. Determine that there are no obstacles around the target position between the starting position and the target position;

(3) Starting to take off, each aircraft moves to its respective target position, and every time a certain aircraft moves one step, scan the global aircraft position;

(4) store the value obtained by the latest scan in step (3), and prepare for the next update;

(5) Calculate the distance between two adjacent planes according to the latest stored data, and judge whether any plane meets the obstacle avoidance condition, that is, whether the closest distance is close to the safe distance;

(6) Entering the state of formation flight, assuming that the obstacle avoidance conditions are met, the model and explanation of the obstacle avoidance process are as follows:

a. Front obstacle model, r is the distance from the outermost periphery of the rotor to the geometric center of the aircraft after the aircraft takes off; R is the distance between the geometric centers of aircraft A and B; assuming that the distance between the two aircraft at a certain time is L, the theoretical safety distance is defined as L > 2r;

方向的速度分量，且满足υ_A＜υ_B，即说明在较短时间内，飞机B会靠近飞机A的安全距离，进而可能发生碰撞；υ _A and υ _B are the planes A and B along the

The speed component of the direction, and satisfy υ _A < υ _B , which means that in a short time, the aircraft B will approach the safe distance of the aircraft A, and then a collision may occur;

方向，速度差为Δυ(t)＝υ_B(t)-υ_A(t)，最大碰撞时间为：

The avoidance scheme under this model is: aircraft B enters the obstacle avoidance state, and adopts the strategy of flying along the gradient direction, so that υ _B decreases: at this time, at

direction, the speed difference is Δυ(t)=υ _B (t)-υ _A (t), and the maximum collision time is:

方向，则只需要保证在t_Δ(t)内Δυ(t)＝0；则飞机A、飞机B这两架飞机不会发生碰撞，全局范围内，实时检测两两飞机之间的距离，其中飞机A与飞机B是某一时刻检测到的最近的两架飞机；According to the requirement to reach the target position in the shortest time, there are two directions of gradient directions in the plane. It is assumed that according to the relevant algorithm, aircraft B has selected

direction, it is only necessary to ensure that Δυ(t)=0 within t _Δ (t); then the two planes, plane A and plane B, will not collide. Globally, the distance between the two planes is detected in real time, where Plane A and Plane B are the two closest planes detected at a certain moment;

方向前进，飞机A与飞机B具有相同轨迹，则此时仍然需要满足上述Δυ(t)＝0的要求，策略改变成飞机B沿着

方向减速；At the same time, according to the relevant algorithm, it is assumed that the aircraft B can only be

If the aircraft A and B have the same trajectory, the above requirement of Δυ(t)=0 still needs to be met at this time, and the strategy is changed so that the aircraft B follows the same trajectory.

direction deceleration;

方向的速度υ_A增大，减小Δυ(t)，相对延长t_Δ(t)，同时沿着某梯度方向，增大与飞机B飞行方向的距离；b. Back-collision model, assuming that aircraft A is required to avoid collision at this time, the avoidance target of aircraft A is not to be hit by aircraft B; first, at this time, aircraft A accelerates so that the

The speed υ _A in the direction increases, decreases Δυ(t), relatively prolongs t _Δ (t), and at the same time, along a certain gradient direction, increases the distance from the flight direction of aircraft B;

Repeat steps (3), (4), and (5), and adopt step (6) for the obstacle avoidance strategy.

The beneficial effects of the present invention are:

1. Better realization of dynamic obstacle avoidance within the fleet and machine-to-machine.

2. It makes up for the deficiencies in the traditional dynamic obstacle avoidance algorithm, and provides an effective reference scheme.

3. An effective reference scheme is given for the complex formation transformation problem in the collaborative formation of multi-UAV groups.

Description of drawings

FIG. 1 is a schematic diagram of an obstacle avoidance model of the present invention.

FIG. 2 is a schematic diagram of the avoidance strategy of the front obstacle model of the present invention.

FIG. 3 is a schematic diagram of the avoidance strategy of the back-collision model of the present invention.

Figure 4 is a schematic flow chart of the present invention.

Detailed ways

The technical solutions of the present invention will be described in further detail below with reference to the accompanying drawings.

As shown in Figure 4, a UAV formation obstacle avoidance method based on gradient function:

(1) Prepare several drones;

(2) Specify the starting position and the target position (pattern);

a. Determine the starting position of the drone fleet;

b. Determine the target location of the drone fleet;

c. Determine that there are no obstacles around the target position between the starting position and the target position;

(3) Starting to take off, each aircraft moves to its respective target position, and every time a certain aircraft moves one step, scan the global aircraft position;

(4) Store the value obtained by the latest scan, and prepare for the next update;

(5) Calculate the safe distance according to the latest stored data, and judge whether any aircraft meets the obstacle avoidance conditions (whether the closest distance is close to the safe distance);

(6) Entering the state of formation flight, assuming that the obstacle avoidance conditions are met, the model and explanation of the obstacle avoidance process are as follows:

a. Figure 1 is the front obstacle model, r is the distance from the outermost periphery of the rotor to the geometric center of the aircraft after the aircraft takes off. R is the distance between the geometric centers of aircraft A and B. Assuming that the distance between the two planes is L at a certain moment, the theoretical safety distance is defined as L>2r.

方向的速度分量，且满足υ_A＜υ_B，即说明在较短时间内，飞机B会靠近飞机A的安全距离，进而可能发生碰撞。υ _A and υ _B are the planes _A and _B along the

The speed component of the direction, and satisfy υ _A <υ _B , which means that in a relatively short time, the aircraft B will approach the safe distance of the aircraft A, and then a collision may occur.

方向，速度差为Δυ(t)＝υ_B(t)-υ_A(t)，最大碰撞时间为：

The avoidance scheme under this model is: aircraft B enters the obstacle avoidance state and adopts the strategy of flying along the gradient direction, as shown in Figure 2, so that υ _B decreases: at this time, at

direction, the speed difference is Δυ(t)=υ _B (t)-υ _A (t), and the maximum collision time is:

方向，则只需要保证在t_Δ(t)内Δυ(t)＝0。则飞机A、飞机B这两架飞机不会发生碰撞(飞机B不会撞向飞机A)。这里必须说明，全局范围内，实时检测两两飞机之间的距离，其中飞机A与飞机B是某一时刻检测到的最近的两架飞机。同时，根据相关算法，假设判定飞机B只能沿

方向前进，则此时仍然需要满足上述Δυ(t)＝0的要求，策略改变成飞机B沿着

方向减速。According to the requirement to reach the target position in the shortest time, there are two gradient directions in the plane. It is assumed that according to the relevant algorithm, aircraft B has selected

direction, it is only necessary to ensure that Δυ(t)=0 within t _Δ (t). Then the two planes, plane A and plane B, will not collide (airplane B will not collide with plane A). It must be explained here that, in the global scope, the distance between two planes is detected in real time, wherein the plane A and the plane B are the two closest planes detected at a certain moment. At the same time, according to the relevant algorithm, it is assumed that the aircraft B can only be

direction, then the above requirement of Δυ(t)=0 still needs to be met at this time, and the strategy is changed to aircraft B along the

direction to slow down.

方向的速度υ_A增大，减小Δυ(t)，相对延长t_Δ(t)，同时沿着某梯度方向，增大与飞机B飞行方向的距离。重复步骤(3)(4)(5)，避障策略采用步骤(6)。b. The schematic diagram of the back-collision model is still shown in Figure 1. Assuming that an avoidance request is made to aircraft A at this time, as shown in Figure 3, the avoidance target of aircraft A is not to be collided by aircraft B. First, at this time, the aircraft A accelerates so that along the

The speed υ _A in the direction increases, decreases Δυ(t), and relatively lengthens t _Δ (t), and at the same time, along a certain gradient direction, increases the distance from the flight direction of aircraft B. Steps (3) (4) (5) are repeated, and the obstacle avoidance strategy adopts step (6).

According to the disclosure and teaching described above, those skilled in the present invention can also make appropriate changes and improvements to the above-mentioned embodiments. Therefore, the present invention is not limited to the above-mentioned specific embodiments disclosed and described. Modifications and changes should also fall within the protection scope of the claims of the present invention. In addition, although some specific terms are used in this specification, these terms are only for convenience of description and do not constitute any limitation to the present invention.

Claims (1)

1. An unmanned aerial vehicle formation obstacle avoidance method based on a gradient function is characterized by comprising the following steps

(1) Preparing a plurality of unmanned aerial vehicles;

(2) specifying a starting position and a target position;

a. determining the starting position of the unmanned aerial vehicle cluster;

b. determining a target position of an unmanned aerial vehicle cluster;

c. determining the position between the starting position and the target position, wherein no obstacle exists around the target position;

(3) starting taking off, moving each airplane to a respective target position, and scanning the position of the global airplane when one airplane moves by one step;

(4) storing the value obtained by the latest scanning in the step (3) and preparing for the next updating;

(5) calculating the distance between every two adjacent airplanes according to the latest stored data, and judging whether any airplane meets an obstacle avoidance condition, namely whether the closest distance is close to a safe distance;

(6) entering a formation flying state, and establishing a model in an obstacle avoidance process on the assumption that obstacle avoidance conditions are met;

repeating the step (3), the step (4) and the step (5), wherein the step (6) is adopted as an obstacle avoidance strategy;

the model established in the obstacle avoidance process in the step (6) is explained as follows:

a. the front barrier model, r is the distance from the outermost periphery of the rotor wing to the geometric center of the airplane after the airplane takes off; r is the distance between the geometric centers of the airplanes A and B; assuming that the distance between two airplanes is L at a certain moment, the theoretical safety distance is defined as L being more than 2 r;

υ _A and upsilon _B Along planes A and B, respectively

A velocity component of direction and satisfies upsilon _A ＜υ _B That is, it means that the aircraft B approaches the safe distance of the aircraft a in a short time, and further collision may occur;

the avoidance scheme under the model is as follows: the airplane B enters an obstacle avoidance state and adopts a strategy of flying along the gradient direction to ensure that upsilon is _B And (3) reducing: at this time is

Direction, velocity difference Δ ν (t) ═ ν _B (t)-υ _A (t), the maximum time to collision is:

there are two directions of gradient direction in the plane, as required to reach the target position in the shortest time, assuming that aircraft B has selected

Direction, then only need to be guaranteed at t _Δ (t) Δ ν (t) ═ 0; the two airplanes A and B do not collide, and the distance between every two airplanes is detected in real time in the global range, wherein the airplane A and the airplane B are the two closest airplanes detected at a certain moment;

at the same time, suppose that aircraft B is determined to be only along

The aircraft A and the aircraft B advance in the same direction and have the same track, the requirement that the delta upsilon (t) is 0 still needs to be met at the moment, and the strategy is changed to the strategy that the aircraft B follows

Decelerating the direction;

b. a rear collision model, wherein if the avoidance requirement is provided for the airplane A at the moment, the avoidance target of the airplane A is not collided by the airplane B; first, the aircraft A accelerates so that it follows

Velocity in direction v _A Increase, decrease Δ ν (t), and extend t relatively _Δ (t) simultaneously increasing the distance from the flight direction of the aircraft B along a certain gradient direction.

Images