改进的粒子滤波在四旋翼姿态估计中的应用

Application of the improved particle filter to quad-rotor aircraft's attitude estimation

针对四旋翼飞行器非线性模型的姿态估计问题,提出了一种基于重要密度函数优选的改进粒子滤波(IDOPF)姿态估计算法。该算法通过扩大重要密度函数的覆盖范围,主动从众多重要密度函数中选择更优初始化粒子群。结合动力学模型和基于反步法控制器的相关控制数据,进行了IDOPF粒子滤波算法在四旋翼飞行器姿态估计中的仿真实验。与EKF算法相比,该算法具有更高的估计收敛速度和估计精度,避免了不稳定滤波,改善了滤波效果,验证了IDOPF算法在四旋翼姿态估计的可行性和有效性。

This paper presents an estimation method based on the important density function optimization particle filter （IDOPF） in order to solve the problem of quad-rotor aircraft＇ s attitude estimation. The IDOPF which depends on the important density function select severity according to expand the scope of important density estimation function, and the more optimization particles is acquired by important density functions. The experiments of simulation were carrying out on the control dataset of dynamic model and backstepping controller. Eventually, the results of experiments show that, compared with EKF, the IDOPF has higher estimation convergence speed and higher accuracy, which avoids the unstable filtering, improves the filtering effect and verifies that IDOPF is feasible and effective in the quad-rotor attitude estimation.