Feasibility Study of Fuel Consumption Prediction Model by Integrating Vehicle-Specific Power and Controller Area Network Bus Technology

Abstract

With the development of automobile technology, it becomes possible to estimate fuel consumption by using the driving parameters from the electronic control unit (ECU) and output the calculation results via the controller area network (CAN) bus. As a result, it is feasible to design an approach for estimating fuel consumption by integrating vehicle-specific power (VSP) and CAN bus technology because of the accessibility and stability of the CAN bus. To determine whether the CAN-based measured data can be used to build the relationship between VSP and fuel consumption to replace the traditional fuel consumption meter, a comparison of fuel consumption data collected from the ECU and the fuel consumption meter is conducted in this study. Results show that the relationship between the fuel consumption rate and VSP bin built by the CAN bus data is consistent with the relationship derived from the fuel consumption meter; this finding indicates that the CAN bus technology can be used to describe the relationship between vehicle activities and fuel consumption rates for light-duty vehicles. In addition, a comparison of CAN-based measured data with VSP-based predicted data shows that the prediction approach that integrates VSP and CAN bus technology needs an aggregation level of 60 s or longer, which thus can be used to estimate a long period of fuel consumption accumulation and a fuel consumption factor for various travel speeds and road types.