摘要
在四旋翼飞行器控制系统中,通常用遗传粒子群算法(GAPSO)来对PID的三个参数进行整定。但是这种方式没有考虑到粒子群算法初期,种群的全局搜索能力强,不需要对粒子进行调节。盲目的在算法中引入交叉、变异操作,只会降低控制系统的效率。因此提出了一种自适应粒子群算法。通过计算种群粒子间距来判断粒子群是否聚集,如果间距小于阈值,就对粒子进行调节来保持粒子群的多样性,否则就不使用。这种调节方式极大地提高了算法的效率,对于需要实时控制的四旋翼飞行器来说以上措施尤为重要。经Matlab/Simulink仿真后结果证实了算法的有效性。
In the quadrotor aircraft control system, genetic particle swarm optimization（GAPSO） is usually used to optimize the three parameters of PID control. But this way does not take into account the facts that the global search ability if particle swarm optimization algorithm is strong at the initial stage and the particle does not need to be adjus- ted. The efficiency of the control system would just be reduced if crossover and mutation operations were introduced blindly in the algorithm. So we propose a self-adapt scatter particle swarm optimization algorithm in this article. In the article, the distance of particle was calculated to determine whether the particle swarm particle spacing gathered themselves together or not, if the distances were smaller than the threshold value, the particle should be adjusted to keep the diversity of particle swarm, otherwise it can not be used. This regulating rule greatly improves the efficiency of the algorithm, and it is very important especially for the quadrotor aircraft control system which needs real-time control. The Matlab/Simulink simulation results confirm the validity of tbe algorithm.
出处
《计算机仿真》
北大核心
2018年第3期29-33,共5页
Computer Simulation
基金
山西省自然科学基金项目(2015011050)
关键词
四旋翼飞行器
自适应粒子群算法
仿真
控制器
Quadrotor aircraft
Self-adapt particle swarm optimization
Simulation
Controller
