Real-time Markov localization for autonomous UGV

Abstract

Localization is the problem of determining the position of a mobile robot from sensor data. The computation of localization algorithm is always expensive and requires high processing power. This paper proposes a special purpose CISC implementation for computing, in real-time, a modified Markov localization algorithm for terrain based navigation control of a light unmanned ground vehicle (UGV). A modified Markov localization algorithm that is capable of overcoming uncertainties resulting from practical, non-uniform, ultrasonic readings is proposed. The proposed design is verified by both simulation and real time results on an experimental non-holomonic robot. Results show the efficiency of the proposed implementation of Markov localization algorithm on the dedicated CISC designed processor.