Probabilistic Uncertainty Identification Modelling in USV Obstacle Detection

R & D Journal of the South African Institution of Mechanical Engineering 2013, 29, 36-43 http://www.saimeche.org.za (open access) © SAIMechE All rights reserved. 36 Uncertainty and false alarms exist in the operational environment of autonomous marine robots. This paper presents the investigations on obstacle detection and avoidance model of an autonomous marine vehicle using ultrasonic sensors and probabilistic techniques in uncertainty identification in the development of an unmanned surface vehicle (USV). Probabilistic techniques were used in modelling the realistic environmental uncertainties of the autonomous marine vehicle. Coherence and non-coherence of sonar signals were investigated and its functions simplified to address the deterministic and un-deterministic modelling variables that were identified in the realistic domain of the USV. The results obtained from the models were adapted into the control algorithm of the USV to reduce the uncertainties present in the guidance and control of the USV.