Mimicking a blind person navigating a corridor using a K-Sonar with a mobile robot

摘要

Our Research Goal is the design of a wall following system based on observations of navigation by blind people with the K-Sonar mobility aid. We seek to gain insight into how blind people navigate and to translate this insight into an autonomous navigation strategy useful to mobile robotics. Current mobile robots that use ultrasonic sensing to follow walls use multiple sensors and rely on assumptions about the direction of echo return. Also they use only range data to the nearest reflecting point, making further assumptions about the reflected echo. In this paper, we describe an advanced wall following algorithm, where the robot follows a wall using a single directed Continuous Transmission Frequency Modulated 'CTFM' Ultrasonic Sensor. The sensor is mechanically panned to track the wall, avoid collisions, and check for navigable space. We further describe an algorithm that reduces the difficulty in ascertaining the bearing of the robot to the wall it tracks. We present results from our mobile robot wall following experiments using range and bearing data from a single directed CTFM ultrasonic sensor. These experiments allowed us to compare our robot sensing and navigation system to the approach we have observed blind people use. The experiments herein described demonstrate the effectiveness of the algorithms we developed from our observations.