Search

Trajectory planning with the involvement of motion time has become a key and challenge for autonomous systems. This paper investigates trajectory planning of unmanned aerial vehicles (UAVs) under maneuverability and collision avoidance constraints. First, a polynomial-based trajectory planning framework is established, and a nonlinear programming problem (NLP) is formulated. Then, a novel asymptotic optimization approach is proposed to improve NLP solution success rate. Three operations of dividing the original NLP into sub-problems, adding constraints gradually, and using previous NLP solution as current initial guess value are designed in the approach. Third, an improved particle swarm optimization (PSO) path planning is also proposed to generate initial guess value for the first sub-problem. Benefited from these operations, the NLP solution success rate is significantly improved. Finally, simulations on simultaneous attack of a same target are carried out. Comparisons with other algorithms illustrate the advantage of the proposed approach. [ABSTRACT FROM AUTHOR]

Automatic generation of optimized flyable path is a key technology and challenge for autonomous unmanned aerial vehicle (UAV) formation system. Aiming to improve the rapidity and optimality of automatic path planner, this paper presents a three dimensional path planning algorithm for UAV formation based on comprehensively improved particle swarm optimization (PSO). In the proposed method, a chaos-based Logistic map is firstly adopted to improve the particle initial distribution. Then, the common used constant acceleration coefficients and maximum velocity are designed to adaptive linear-varying ones, which adjusts to the optimization process and meanwhile improves solution optimality. Besides, a mutation strategy that undesired particles are replaced by those desired ones is also proposed and the algorithm convergence speed is accelerated. Theoretically, the comprehensively improved PSO not only speeds up the convergence but also improves the solution optimality. Finally, Monte-Carlo simulation for UAV formation under terrain and threat constraints are carried out and the results illustrate the rapidity and optimality of the proposed method. • A novel and comprehensively improved PSO is proposed and analyzed. • Path planning for UAV formation is evaluated by Monte-Carlo simulations. • Faster convergence speed and better solution optimality are achieved. [ABSTRACT FROM AUTHOR]