基于改进雁群PSO算法的模糊自适应扩展卡尔曼滤波的SLAM算法

Fuzzy adaptive extended Kalman filter SLAM algorithm based on the improved wild geese PSO algorithm

针对扩展卡尔曼滤波同时定位与地图创建算法中难以建立准确的先验噪声模型的问题,提出一种基于改进雁群粒子群算法的模糊自适应卡尔曼滤波算法.利用分数阶微积分改进粒子进化速度,利用混沌来改进粒子的初始化和发生早熟时的处理.改进后的雁群粒子群算法在收敛速度与避免早熟方面有了很大改进,并将改进的雁群粒子群算法用于模糊自适应扩展卡尔曼滤波同时定位与地图创建算法的训练,并与用雁群粒子群算法训练的模糊自适应扩展卡尔曼滤波同时定位与地图创建算法进行对比,其在定位与构图方面有很大的提高.

Because the problem of the extended Kalman filter localization and mapping algorithm priori noise model is difficult to manage, this paper proposes an improved wild geese particle swarm algorithm based on the fuzzy adaptive Kalman filter localization and mapping algorithm. We take advantage of the the fractional calculus to improve particle speed of evolution, and make use of chaos to improve the initialization of the particle and the precocious one when processing. The improvement of wild geese particle swarm algorithm is shown in convergence rate and avoiding premature, then they can improve geese particle swarm algorithm for fuzzy adaptive extended Kalman filter localization and mapping algorithm training, in contrast with geese particle swarm algorithm fuzzy adaptive extended Kalman filter simultaneous localization and mapping algorithm, the new algorithm positioning and composition has greatly improved.