City Bus Mass Estimation based on GPS and CAN-Collected Tracking Data

This paper deals with estimation of city bus mass based on regularly, on-line collected engine CAN data and vehicle tracking GPS data. Since mass estimation based on instantaneous recording is not usable due to imprecise inputs and, correspondingly, large variance of estimates, an alternative approach is employed based on using multiple consecutive data recordings for the interval between two subsequent bus stations. The estimation methodology is based on bus longitudinal dynamics model and least squares method. The key characteristic, which needs to be properly reconstructed for accurate mass estimation, is related to dependence of total wheel power on measured engine power load. The proposed estimation method is parameterized by using the data recorded on a real bus route in the city of Dubrovnik, and validated by using passenger mass measurement by means of passenger counting.