Development of a Realistic Driving Cycle Using Time Series Clustering Technique for Buses: Thailand Case Study

Realistic driving cycles for Thailand's road conditions were studied only for the pollution problems of passenger cars and light-duty trucks in urban areas of the metropolitan region. Such driving cycles did not cover rural areas and the in-use operation of buses. Furthermore, such created methods of the driving pattern were very irregular and complicated for chassis dynamometer operation. To propose a new method for a realistic driving cycle, the development of the route traffic for bus transportation in rural areas using a time series clustering technique is indicated. As a case study, this method was applied to collect the driving data on the 323 route in Kanchanaburi province using on-board measurement. The selection procedure of suitable speed and interval time for driving cycle construction was revealed. Similarity of driving characteristics was identified with clustering technique for each time duration to decide the best driving cycle. In conclusion, the frequency of speed ranges from entire trips is 30-40 km/h with the highest ratio of deceleration time to the entire trip. Furthermore, discrete average speeds at each time point with 40 seconds of interval time are the best choice related to the real driving condition in this case study.