Terrain Feature Extraction and Classification for Mobile Robots Utilizing Contact Sensors on Rough Terrain

Abstract

To run safely and carry out the mission, a mobile robot should identify types, physical and/or geometric characteristics of terrains on which the robot runs. In this paper, we designed and implemented a mobile robot for collecting contact sensor data that come from the reaction between ground and tires, and then collect sensor data from four experimental terrains. After analyzing the collected data and the physical reaction, we suggested a new method for terrain feature extraction and terrain identification. Also we compared a conventional method for terrain feature extraction using Fast Fourier Transform(FFT) with the proposed method and then we concluded the conclusion that our method is more efficient than the other existing method.