基于引力场优化的Unscented FastSLAM2.0算法

为了改进Unscented Fast SLAM2.0算法重采样过程中的"粒子退化"和"粒子贫化"问题,本文提出了一种基于引力场优化的Unscented Fast SLAM2.0算法.首先采用Unscented粒子滤波器替代扩展卡尔曼滤波估计移动机器人路径后验概率,然后采用扩展卡尔曼滤波器对环境进行估计更新,最后用引力场优化思想优化重采样过程:在重采样中每个采样粒子近似成宇宙灰尘,通过引力场的移动因子产生作用驱动粒子集更快朝着真实的机器人位姿状态逼近,改善粒子退化问题:通过自转因子的自转作用,避免粒子过分集中,保障了粒子多样性.实验结果表明了该算法的有效性.

A unscented fastslam2.0 algorithm based on gravitational field optimization

To solve the problem of particle degradation and sample dilution in Unscented FastSLAM2.0 algorithm, this paper proposed a Unscented FastSLAM2.0 algorithm based on gravitational field optimization (GFA-Unscented FastSLAM2.0).Firstly, using Unscented particle filter instead of the Extended Kalman filter (EKF) to estimate the posterior probability of mobile robot path.Then the Extended Kalman filter is used to estimate and update the environment.Finally the gravitational field algorithm is used to optimize the resampling proces:each particle is considered cosmic dust during resampling proces, throung the mobile gravity factor of Gravitational Field can drive the particle swarm towards the real robot position faster, and improve the particle degradation problem; through the rotation effect of rotation factors can avoid excessive concentration particles, so as to ensure the diversity of particles.The simulation results show the effectiveness of the new algorithm.