基于粒子滤波的未知环境下机器人同时定位、地图构建与目标跟踪

为了解决机器人在未知环境下的目标跟踪问题，提出了一种基于粒子滤波的机器人同时定位、地图构建与目标跟踪方法．该方法采用Rao-Blackwellized粒子滤波器对机器人位姿状态、标志柱分布和目标位置同时进行估计．该方法中，粒子群的总体分布情况表征机器人位姿状态，而每个粒子均包含2类EKF滤波器，其中一类用来完成对标志柱分布的估计，另一类用来完成对目标状态的估计，粒子的权值则由粒子状态相对于标志柱和目标状态2类相似度共同产生．通过仿真和实体机器人实验验证了该方法的有效性．

Simultaneous localization,mapping and object tracking in an unknown environment using particle filtering

The proposed research paper examines a simultaneous localization, mapping, and object tracking method. The examination was in part based on a particle filter that allows a robot to track an object in an unknown environment. This method utilizies the Rao-Blackwellized particle filting to estimate the pose of robot, landmarks distribution, and object position simultaneously. The general distribution of a particle swarm represents the pose of a robot, and each particle includes two kinds of Extended Kalman Filter (EKF). One EKF estimates distribution of landmarks, while the other EKF estimates the state of the object. The weight of particle is determined by the combination of two likelihoods, one is the likelihood between particle state and landmarks, and the other is the likelihood between particle state and object state. The results of the research indicate the valid robot experimentation and simulation, confirm the proposed research approach is very effective.