论文部分内容阅读
为解决四旋翼飞行器飞行控制中的稳定性问题,尤其是姿态解算精确性问题,提出改进EKF算法。该算法采用两个EKF并行地对四旋翼飞行器的姿态数据进行处理,将改进的EKF输出的状态变量和协方差加权后整合为飞行器的输出状态和协方差对飞行器姿态进行实时解算。该算法有效地克服了非高斯白噪声对飞行器姿态解算的影响,减小了姿态数据的滤波偏差。仿真表明,该算法较之普通EKF算法的滤波效果更好,改进EKF的均方误差较之普通EKF降低了43.8%,提高了系统的鲁棒性。最后搭建以NI my RIO为核心控制器的四旋翼飞行器,验证改进EKF算法在四旋翼飞行器系统上的正确性和有效性,且该算法亦能满足飞行器实时控制的需求。
In order to solve the stability problem of four-rotor aircraft flight control, especially the accuracy of attitude solution, an improved EKF algorithm is proposed. The algorithm uses two EKFs to process the attitude data of quadrotors in parallel, and integrates the state variables and covariances of the improved EKF outputs into the output states and covariances of the aircraft to solve the attitude of the aircraft in real time. The algorithm effectively overcomes the influence of non-Gaussian white noise on the attitude of the aircraft and reduces the filtering deviation of the attitude data. The simulation results show that this algorithm has better filtering performance than the ordinary EKF algorithm, and the mean square error of the improved EKF is reduced by 43.8% compared with the normal EKF, which improves the system robustness. Finally, the quadrotorcraft with NI my RIO as the core controller is established to verify the correctness and validity of the EKF algorithm in quadrotor system. The proposed algorithm can also meet the real-time control requirements of the aircraft.