基于多次测量更新的移动机器人SLAM仿真  

Mobile Robot SLAM Simulation with Multi Measurement Update

作　　者：许亚芳 孙作雷[1] 曾连荪[1] 张波[2] 

机构地区：[1]上海海事大学信息工程学院,上海201306 [2]中国科学院上海高等研究所,上海201210

出　　处：《系统仿真学报》2015年第6期1288-1293,共6页Journal of System Simulation

基　　金：国家自然科学基金项目(61105097;51279098;6140127);上海市教育委员会科研创新项目(13YZ081)

摘　　要：针对移动机器人同时定位与地图构建(SLAM)技术中,非线性系统进行线性化处理所产生的线性化误差问题,从费切尔信息角度分析,提出多次测量更新算法。该算法利用滤波中预测和各次更新阶段状态向量的费切尔信息加权关系,解算每次测量更新后的状态估计,通过多次将更接近于真实状态的估计值融入观测信息,可得到较高精度的后验状态估计,降低线性化误差的同时提高了机器人定位与构图准确性。实验部分实现多次测量更新与一次测量更新算法的仿真与对比,结果表明:所提算法能有效降低机器人位姿误差和地图估计误差。Aiming at the problem of the accumulation of linearization error in the nonlinear system linearizing of Simultaneous Localization and Mapping(SLAM) in mobile robot,an algorithm named multi measurement update was put forward according to the analysis of Fisher information.In order to compute the state estimation after each measurement update,the Fisher information weight relationship between prediction variable and update variable was made use of.Due to a number of data association with an estimation which was more close to the real data than the former,the algorithm could achieve a more accuracy posterior state.As a result,it could decrease the linearization error and improving the precision of localization and mapping.The experiments made a comparison between the multi measurement update and single measurement update.It shows that the proposed method can efficiently reduce the robot pose error and map information error.

关 键 词：移动机器人 费切尔信息 多次测量更新 线性化误差 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]