Dual-sensor fusion for obstacle avoidance in indoor environment

This paper presents a real-time automatic obstacle avoidance system using depth information and sonar data. The depth map is provided by Microsoft Kinect and sonar data are obtained from sonar sensors. With fusion of weighted depth information and sonar data, a set of intelligent fuzzy rules are designed to construct a safe path to avoid obstacles in indoor environment. According to the experimental results, this system has good performance while compared with the previous approaches and can work in dark environment. According to the experimental results, this system has good performance while compared with the previous approaches and can work in unfamiliar environment.