摘要
针对目前FastSLAM改进算法存在复杂度高的问题,提出一种基于差分粒子滤波器(DDPF)-差分滤波器(DDF)的同时定位与地图创建方法。该方法采用DDPF估计机器人路径,采用DDF估计陆标位置,同时采用可选重采样以降低样本贫化的概率。实验结果表明,该方法具有精度高、连贯性好以及复杂度适中的特点。
A simultaneous localization and mapping method was brought forward based on Divided Difference Particle Filter-Divided Difference Filter （DDPF-DDF） in allusion to the high complexity problem of FastSLAM improvement algorithm. This method adopted DDPF to estimate the path of robot and DDF to estimate the landmark position. It also adopted selectable resample to lower the probability of sample deletion. The experimental results indicate that the method has high precision,good consistency and moderate complexity.
出处
《计算机应用》
CSCD
北大核心
2009年第11期3060-3063,3067,共5页
journal of Computer Applications
关键词
FASTSLAM算法
差分粒子滤波器
差分滤波器
同时定位与地图创建
重采样
FastSLAM algorithm
Divided Difference Particle Filter （DDPF）
Divided Difference Filter （DDF）
simultaneous localization and mapping
resample
