Research on Passenger Car Driving Cycle with Multi-Source Data Fusion: Identification, Classification and Signal De-noising 2020-01-1044
Drivability plays a decisive role in evaluating vehicle performance, which directly affects the willingness of drivers and passengers to consume. In order to overcome the shortcomings of subjective evaluation due to expensive and time consuming. An objective evaluation system for passenger car driving cycle with multi-source data fusion is developed, which can be used to implement parameters acquisition, cycle identification, signal de-noising and feature value extraction. In this paper, recognition and classification of vehicle driving cycle and de-noising of acceleration signal are focused. First, the main parameters for objective driving performance evaluation are obtained by analyzing the vehicle transmission system model and the longitudinal dynamics theory. Then, the characteristics of the objective parameters of the specific working conditions and the expert knowledge base are combined, the sliding window method is adopted, and the identification and classification of the mixed working conditions are realized by the mixed programming of Labview and Matlab Software. Then, according to the characteristics of the relevant parameters of objective drivability, the advantages and disadvantages of different filtering methods for signal de-noising and reproduction are analyzed, including high-pass filtering, low-pass filtering, band-pass filtering, wavelet filtering and moving average filtering. Finally, the reliability and practicability of condition recognition and signal de-noising are verified by the combination of Tip-in and Tip-out.
This research can be used as an important reference for the research of vehicle driving performance of hybrid vehicles, electric vehicles and smart cars, and also guides the design of transmission system, parameter adjustment and the method selection of signal de-noising.