A multisensor fusing system on ultrasonic sensors

Abstract

In the application of an industrial robot, it is often found that the workspace is restructured or changed from time to time, as with for example mining robots or some construction robots. In some special cases, where for example there is dust or insufficient illumination in the environment, the use of sensors is limited. Ultrasonic sensors have been widely used in the robot application for its simplicity and low cost. However, determining accurate distance and direction of the obstacle from ultrasonic sensors is quite difficult due to the expanding beam of the ultrasonic waves. The uncertainty about the distance and direction information limits the further application of ultrasonic sensors. In order to use ultrasonic sensors efficiently, a blackboard strategy for multiple ultrasonic sensors is developed in this paper. Based on the blackboard, information of the inner sensors of robot and information from ultrasonic sensors are integrated and fused at the data-level, character level and decision level. Finally, a simulation shows that the uncertainty of the distance and direction information has been eliminated to some degree and a more accurate workspace has been obtained.