Fast Gaussian particle filter data fusion method based on artificial fish swarm optimization

Abstract

The invention provides a fast Gaussian particle filter data fusion method based on artificial fish swarm optimization, belongs to the technical field of signal processing, and is mainly used for solving the problems of huge calculation workload and low precision of a particle filter in a multi-particle state. According to the method, Gaussian particle filtering is used as a framework, an artificial fish swarm algorithm is fused, and foraging behaviors and clustering behaviors are used for optimizing weights. According to the method, traditional sampling is replaced by linear transformation, the weight is optimized according to the measurement value and the weight calculation formula, the calculation speed is guaranteed while the calculation precision is improved, and the method is suitablefor application occasions such as state estimation of a nonlinear dynamic system.

Description

technical field [0001] A fast Gaussian particle filter data fusion method based on artificial fish swarm optimization proposed by the invention belongs to the technical field of signal processing and relates to nonlinear filtering. The method provided by the invention is suitable for state estimation of nonlinear dynamic systems. Background technique [0002] Nonlinear filtering problems arise in many areas, including object tracking, strapdown inertial navigation systems, and attitude estimation. Extended Kalman Filter (Extended Kalman Filter, EKF) is to linearize the nonlinear function and directly truncate the high-order items, resulting in large errors and low filtering accuracy. Unscented Kalman Filter (UKF) is a Kalman filter using unscented transformation. Compared with EKF, its filtering accuracy has been improved, but its nonlinear transfer error always exists. Since the extended Kalman filter and the unscented Kalman filter are both based on the improvement of the...

AI Technical Summary

Problems solved by technology

Particle Filter (PF) is a filtering algorithm based on the Monte Carlo MONTE-CARLO method, but the resampling strategy in the particle filter algorithm will directly affect the performance of the filter, which will affect the real-time performance of the algorithm.

To this end, an improved Gaussian particle filter algorithm (Gaussian Particle Filter, GPF) is proposed, which approximates the posterior distribution of unknown variables through Gaussian distribution, without resampling, which greatly saves filtering time, and is better than particle filter in real-time The performance of the algorithm is better than that of EKF, UKF and other algorithms. In order to improve the running speed, a linear Gaussian particle filter algorithm is proposed, that is, it is replaced by linear transformation when sampling. Compared with the particle filter, the accuracy is greatly shortened. Running time, but there are limitations in use, and there is room for improvement in accuracy

In order to continue to improve the filtering accuracy, a particle filter method based on artificial fish swarm is proposed. This algorithm adds the optimization of artificial fish swarm to the particle filter algorithm, which improves the filtering accuracy to a certain extent, but there are still improvements in the calculation speed and accuracy. space

Images