GPS latency identification by Kalman filtering

For outdoor mobile robots localisation, GPS has obvious advantages: position data are directly given in an absolute frame, and the required infrastructure is reduced to a sole fixed station in the case of differential systems. Yet, the use of this solution raises a number of issues, such as the satellite maskings, or the existence of the so-called GPS latency which delays the output of the localisation data. This paper deals with the latter problem, and proposes a method to identify this parameter without using other exteroceptive sensors. Experimental results are presented based on trials carried out at the Centre d'Experimentations Routieres (CER) in Rouen, in collaboration with the Laboratoire Central des Fonts et Chaussees (LCPC).