Fusion of multiple ultrasonic sensor data and imagery data for measuring moving obstacle's motion

Abstract

Two different functions, intensive and wide-angle observations, are required for robot vision. In this research, multiple ultrasonic sensors are selected for wide-angle observation, and an image sensor for the intensive observation, and methods of measuring moving obstacle motion are proposed by two kinds of fusion: 1) fusion of the two different sensor data, and 2) fusion of multiple ultrasonic sensor data. The latter fusion methods utilize the movement of the obstacle from a measuring range of an ultrasonic sensor to other sensor range. They are formulated in the framework of Kalman filter. Simulations and experiments show the effectiveness and applicability to a real robot system. Additionally, the proposed method can be estimated as a new framework of sensor fusion that the fusion is performed by the movement of the object from one sensor range to others.