Performance evaluation of proprioceptive sensor sets for civil-engineering articulated vehicle localisation

In order to localise civil-engineering vehicles, an extended Kalman filter is used, which combines the exteroceptive data given by a real-time kinematic GPS with proprioceptive measures issued by internal sensors. The main problem of this application is that the potential users of the system are reluctant to use additional encoder wheels. This paper proposes different sets of internal sensors and compares their respective performances so as to determine the best (and cheapest) alternative to odometry. Experimental results are presented from data obtained during trials carried out on an instrumented compactor.