【摘要】 在开源框架Player平台的基础上搭建了针对室内服务机器人的地图构建和定位系统。首先将DP-SLAM算法移植到Player,离线构建动态地图,以减少人工绘制地图的误差和制约;然后引入一种KLD取样的适应的蒙特卡洛(KLD-AMCL)定位算法,通过计算MLE与真实后验的KL距离,自适应调整所需粒子数;最后结合Player平台、动态地图和KLD-AMCL算法,搭建了针对室内服务机器人的定位系统。实验表明,该系统具有较好的环境适应性和较高的定位精度。更多还原

【Abstract】 An indoor service robot map building and positioning system is built based on the open source platform Player.First,DP-SLAM algorithm is transplanted to the Player to build the dynamic maps offline,in order to reduce errors and constraints caused by manual map building.Then,the KLD-Sampling Adaptive Monte Carlo locating(KLD-AMCL) algorithm is introduced,which can adaptively adjust the required number of particles by calculating the MLE and real posterior KL distance.Finally,an indoor service robot position system is built by combine the Player platform,dynamic map building and KLD-AMCL algorithm.Empirical results show that the system has good environmental adaptability and high positioning accuracy.更多还原