论文部分内容阅读
提出了一种快速估计线性非线性状态的粒子滤波器。针对混合线性非线性模型中传统的Rao-B lackwellized Particle Filter(RBPF)方法对状态估计处理速度较慢的不足,将条件相关的线性与非线性状态进行分离,对于非线性状态运用粒子滤波(PF)算法进行估计,取出从非线性状态中抽取出的粒子的均值,将此均值传播到线性状态中做一次卡尔曼滤波,得出对线性状态的估计。仿真结果表明:与RBPF相比可以将处理速度提高50%~60%,有效改善状态估计的实时性。
A particle filter for fast estimation of linear non-linear state is proposed. Aiming at the problem that the traditional Rao-B lackwellized Particle Filter (RBPF) method is slow to deal with the state estimation in the mixed nonlinear model, the conditional linearity and nonlinearity are separated. For the nonlinear state, the particle filter ) Algorithm, the average value of the particles extracted from the non-linear state is extracted, and the average value is propagated to a linear state to perform a Kalman filter to obtain an estimate of the linear state. Simulation results show that compared with RBPF, the processing speed can be increased by 50% ~ 60%, which can effectively improve the real-time performance of state estimation.