Comparative Estimation of Electric Vehicle Rolling Resistance Coefficient in Winter Conditions

The electric/hybrid vehicle are promising technologies and practical for the transport and the environment. For energy management, the vehicle longitudinal dynamics used to estimate vehicle power demands depend on several parameters, including its mass and its rolling conditions. This study compared the methods of estimating the rolling resistance coefficient and mass to establish energy planning system of an electric vehicle in winter conditions. Different from reported approaches, which are limited to estimate only one parameter, the vehicle mass or the rolling resistance, this paper used two efficient methods which, simultaneously estimate vehicle mass and rolling resistance coefficient. The first method is based on RLS algorithm, while the second is based on neural network. The estimated values of rolling resistance coefficient retrieved from these methods are similar to those provided by a third method, which estimates only the rolling resistance by considering the mass as input and using RLS algorithm. Although results retrieved from the three methods show that the estimated values converge to real values with a margin error that does not exceed 10%, we suggest that the first and the third method, using the RLS algorithm and giving an online estimation, are more accurate and more suitable in snow covered road conditions.