Global Positioning Systems (GPS) provide an effective means of outdoor localisation. Unfortunately they are subject to a variety of errors, particularly in cluttered environments where GPS signal is not always available. Whilst GPS positional information includes measures of signal quality, these can not always be trusted, particularly just after a lost positional fix is regained. This paper presents the use of an Extended Kalman Filter to fuse GPS and odometric measurements, in order to improve vehicle positional and heading estimates. An odometric motion model is used to predict future positions, which are corrected by GPS measurements. Uncertainty in positional information from both GPS and odometry systems is modelled. The system has been implemented on Seekur, a terrestrial platform manufactured by Mobile Robots. Results of an experimental excursion of over 2 km are presented and show the efficacy of the system.
Reference:
Burke, M.G. and Sabatta, D. 2009. Position fusion for an outdoor mobile robot. 3rd Robotics and Mechatronics Symposium (ROBMECH 2009). Pretoria, South Africa, 8-10 November 2009, pp 5
Burke, M. G., & Sabatta, D. (2009). Position fusion for an outdoor mobile robot. http://hdl.handle.net/10204/3817
Burke, Michael G, and D Sabatta. "Position fusion for an outdoor mobile robot." (2009): http://hdl.handle.net/10204/3817
Burke MG, Sabatta D, Position fusion for an outdoor mobile robot; 2009. http://hdl.handle.net/10204/3817 .
