Vehicle localisation using asphalt embedded magnetometer sensors

While advanced driving assistance systems for vehicles are becoming mature in urban and freeway scenarios, the enabling service of precise localisation with respect to the road lane is gaining more and more attention in the last decade. Indeed, having knowledge of the precise location of the vehicle allows safer and more efficient solutions, especially when those information are shared among multiple vehicles. This manuscript deals with the definition of an effective and simple nonlinear least squares solution for estimation and tracking of vehicles using multiple magnetometers embedded on the lateral road stripes. The solution relies on a classical model of magnetometer measurements and it is made robust by adding turning the classic least square solution to a constrained optimisation problem dictated by known road quantities (i.e., curvature and lanes dimensions). A first set of results sound promising for an effective application of the proposed algorithm to vehicle tracking.