APPLICATION OF A NEW ANN-BASED MULTI-SENSOR DATA FUSION ALGORITHM TO OBSTACLE CATEGORY IDENTIFICATION

This paper presents a new multi-sensor data fusion algorithm to identify obstacle categories for an autonomous mobile robot. The 2D image from CCD camera and the range information from an ultrasonic ranging system are fused by two kinds of neural networks-a Cerebella Model Articulation Controller (CMAC) and a multi-layer feedforward network. They are off line trained by a set of data which are collected around the obstacles. To show the effectiveness of the proposed system, a series of simulation experiments were conducted. The results show that the category of an obstacle can be identified in real time using a personal computer.