Online vehicle mass estimation using recursive least squares and supervisory data extraction

This paper examines the online estimation of onroad vehicles´ mass. It classifies existing estimators based on the dynamics they use for estimation and whether they are event-seeking or averaging. It then proposes an algorithm comparable to this literature in accuracy and speed, but unique in its minimal instrumentation needs and ability to provide conservative mass error estimates, in the 3sigma sense. The algorithm builds on the simple idea, inspired by perturbation theory, that inertial dynamics dominate vehicle motion over certain types of maneuvers. A supervisory algorithm searches for those maneuvers, and feeds the resulting filtered data into a recursive least squares-based mass estimator and conservative mass error estimator. Both simulation and field data demonstrate the viability of the resulting approach.