Sonar echo-location in 2-D using mini-microphone array and spatiotemporal frequency filtering

Sonar array technology enables sensing of both range and angle bearing of objects with respect to the array. Instead of relying on a narrow angle of transmission and reception and multiple sensor positions, one array can receive returns from a large area and still offer angular localization of targets. We have designed and constructed array processing circuitry that uses mixed-signal spatiotemporal frequency filtering in order to extract wavefront velocity across the array. From wavefront velocity, the angular bearing of a return is determined. This system was demonstrated in a functional test using single and multiple objects. The 2D performance is presented here.

I. Introduction

Sonar technology is well-known in the mobile robot community for being a low-cost method of obtaining range data about features in the environment. It is unfortunately known chiefly for its economy, due to some well-known drawbacks of plain single-sensor sonar. A single sonar sensor must have a narrow angle of reception, to reduce the confusion of multi-path echoes which return after multiple reflections instead of directly back. A single sonar sensor, unfortunately, can not be sufficiently focused to completely avoid multi-path interference without a large physical acoustic structure. Additionally, to sense a wide area the sensor must either be physically rotated, introducing another mechanical system and more complexity, or there must be many multiple sensors. Many strategies for incorporating the range data from multiple sensors have been proposed [1] [2] [3], relying purely on multiple distance measurements, and processing to discriminate “real” objects from erroneous “phantom” returns.