Abstract
Underwater robot localisation is challenging as it cannot rely on sensors such as the GPS due to electromagnetic wave attenuation or optical cameras due to water turbidity. SONARs are immune to these issues, hence they are used as alternatives for underwater navigation despite lower spatial and temporal resolution. Single-beam SONARs are sensors whose main output is distance. When combined with a filtering algorithm like the Kalman filter, these distance readings can correct localisation data obtained by inertial measurement units. Compared to multi-beam imaging SONARs, the single-beam SONARs are inexpensive to integrate into underwater robots. Therefore, this study aims to develop a low-cost localisation solution utilizing single-beam SONARs and pressure-based depth sensors to correct dead-reckoning linear localisation data using Kalman filters. From experiments, a single-beam SONAR per degree of freedom was able to correct localisation data, without the need of complicated data fusion methods.