Efficient path planning for UAV formation via comprehensivelyimproved particle swarm optimization

Relevant information:

Efficient path planning for UAV formation via comprehensively improved particle swarm optimization - ScienceDirect

Summary:

Automatically generating optimal flight paths is a key technology and challenge for autonomous UAV formation systems. In order to improve the speed and optimality of UAV formation trajectory planning, a three-dimensional trajectory planning algorithm for UAV formation based on the comprehensive improved particle swarm algorithm is proposed. In this method, chaos-based Logistic mapping is first used to improve the initial distribution of particles. Then, the commonly used constant acceleration coefficient and maximum velocity coefficient are designed to adaptively change linearly to adapt to the optimization process and improve the optimality of the solution. In addition, a mutation strategy is proposed to replace undesired particles with desired particles, which speeds up the convergence of the algorithm. Theoretically, the comprehensively improved particle swarm algorithm not only speeds up the convergence speed, but also improves the optimality of the solution. Finally, Monte Carlo simulation was performed on the UAV formation under terrain and threat constraints, and the simulation results verified the rapidity and optimality of the method.

1. Background of text selection

The research background of the subject requires rapid generation of formations.

2. Article adoption points

1. Mountain and radar modeling

Mountain modeling: $z_k^m=h_k*\exp\left(\frac{\left(x_k^m-x_k^{m0}\right)^2}{x_k^t}+\frac{\left(y_k^m-y_k^{m0}\right)^2}{y_k^t}\right)$

Radar modeling: $T_k=\left(x_k^r,y_k^r,z_k^r,R_k\right)$

2. Objective function design

Searching feasible tracks for UAV formation flight is a complex multi-objective optimization problem. Taking into account factors such as path length, environmental constraints and collision avoidance, the evaluation function of UAV formation path planning can be written as

Function design: f = fL+ fT+ fR+ fC

3. Adaptive maximum speed coefficient

In addition to the inertial weight, the acceleration coefficients c1 and c2 are two other key parameters. They indicate the weight of the random acceleration term, which pulls each particle to the local and global optimal position, and plays an important role in adjusting the convergence speed and search direction. In the first half of the phase, the search process dominates, while in the second half, the convergence process dominates. Therefore, the acceleration coefficients c1 and c2 are adaptively designed as:

$\begin{aligned}c_1&=c_{\max}-\frac{(c_{\max}-c_{\min})t}{T}\\c_2&=c_{\min}+\frac{(c_{\max}-c_{\min})t}{T}\end{aligned}$

Where T is the total number of iterations, t is the current number of iterations, cmax and cmin are constant values, where cmax > cmin > 0. In the simulation, c1 and c2 of the traditional PSO are set to the same constant values as in the study,