A Physically Based Navigation Strategy for Sonar-Guided Vehicles

Abstract

A novel navigation strategy is described that employs a time- of-flight sonar system to guide an autonomous vehicle through an unstructured environment composed of specular surfaces. Collisions are avoided by employing a scanning procedure that exploits the physics of sound propagation to detect objects. Physical models for the transducer beam and the reflection process are described. These models provide understanding about sonar ranging and are applied to deter mine the necessary scanning procedure for obstacle detection. In our environment, two types of echoes are observed: those reflected from surfaces and those diffracted from edges. Acoustic pulses reflected from oblique surfaces are not usually detected, and diffraction echoes are small and decrease with distance, making them difficult to detect. A scanning proce dure is developed to detect both types of obstacles. This pro cedure is employed to develop a navigation strategy that indicates the perceived obstacle-free region that guarantees that no collision will occur. The approach is illustrated with results produced by a vehicular robot equipped with a Polaroid sensor.