Abstract: In order to construct a reasonable typical driving cycle of passenger cars, an arithmetic that based on Markov chain principal was proposed and the driving features were investigated based on driving data in the city of Dalian by adopting cluster analysis. The typical driving cycle of passenger cars was constructed and comparison analysis of driving features between the constructed typical Dalian driving cycle and the existing available driving cycles was applied. Results show that the average absolute deviation of driving characteristic parameters between the constructed driving cycle and driving database population is 4.42%. This value indicates that the development requirement that the deviation must be under 5% is met and the validness and accuracy of the constructed arithmetic is verified. Remarkable region discrepancy of driving characteristics of passenger cars exists in domestic cities. NEDC (New European Driving Cycle) can hardly characterize the domestic driving characteristics correctly. Thus, it is necessary to construct driving cycle which can correctly reflect the domestic actual driving features. The results can provide beneficial references for arithmetic of constructing driving cycle and analysis of driving characteristics.

Key words: driving cycle, passenger car, cluster analysis, Markov chain