Human tracking and following of mobile robot with a laser scanner

Abstract

The mobile robot with a capability of following human targets can be used in various ways in different environments. The main technologies of human following robot system contain human detection, trajectory tracking and following control. In this paper, we proposed a method to detect a person's legs by a single laser scanner, and applied random forest classifier to distinguish the legs from other objects. According to the laser scanner data and detection results, the target person with moving speed and relative to the position of the robot will be tracked through Extended Kalman filter (EKF) algorithm to get its moving trajectory. The following control algorithm used the continuous position of human as input and the velocity planning of mobile robot as output to implement the human following function. This method was experimented in office and long corridor, the results showed that the detection and tracking algorithms were accurate and robust, the mobile robot followed the target person correctly and smoothly. All of these functions had been implemented in the Robot Operating System (ROS) framework.