Detection and Tracking of Moving Objects for Indoor Mobile Robots with a Low-Cost Laser Scanner

In this work, we present a detection and tracking of moving object (DATMO) system with a low-cost rotary laser range finder. This system is designed for indoor mobile robots. An occlusion and noise detection module is developed for processing laser range finder’s data. A cascade classifier is proposed for object detection. Then we transform the local target position into the global map using a prior occupancy grid map. An Extended Kalman Filter is applied for object tracking and error compensation. Our system runs at 8 Hz on a Raspberry Pi 3 using a LS01C 2D laser range finder. Indoor human tracking experiments are designed to verify the efficiencies of the algorithms. The presented method was proven to be capable of providing a smooth and accurate target position in a DATMO task.