Optimised trajectory tracking control for quadrotors based on an improved beetle antennae search algorithm

This paper focuses on the trajectory tracking of quadrotors under bounded external disturbances.An optimised robust controller is proposed to drive the position and attitude ofa quadrotor converge to their references quickly. At first, nonsingular fast terminal slidingmode control is developed, which can guarantee not only the stability but also finite-timeconvergence of the closed-loop system. As the parameters of the designed controllers playa vital role for control performance, an improved beetle antennae search algorithm is proposedto optimise them. By employing the historical information of the beetle’s antennaeand dynamically updating the step size as well as the range of its searching, the optimisingis accelerated considerably to ensure the efficiency of the quadrotor control. The superiorityof the proposed control scheme is demonstrated by simulation experiments, from whichone can see that both the error and the overshooting of the trajectory tracking are reducedeffectively.

改进的神经网络PID在空调温度控制中的应用

为提高候机楼中央空调温度控制水平,针对候机楼中央空调系统具有时滞性、扰动因素较多等特点,提出了一种基于改进天牛须搜索(IBAS,improved beetle antennae search)算法的模糊径向基函数(RBF,radial basis function)神经网络(PID,proportion integration differentiation)控制方法,建立了空调区域温度控制模型,通过模糊RBF神经网络实现PID参数在线整定,解决系统非线性、时变的问题。同时由于神经网络参数存在难以选取问题,提出利用天牛须搜索(BAS,beetle antennae search)算法优化模糊RBF神经网络参数的方法,并引入莱维飞行机制和变步长策略对BAS算法进行改进,提高其跳出局部最优的能力和稳定性。仿真结果表明,采用IBAS算法优化的模糊RBF神经网络PID控制方法有效提高了系统的鲁棒性和自适应能力,对候机楼中央空调系统具有良好的控制效果。