Vehicle Emission Characteristic Parameters-based Approach To The Development And Evaluation Of Driving Cycles

The vehicle exhaust emission problem has continuously worsened, which has become the research focus in the area of traffic environment. The emerging and development of the portable emissions measurement system (PEMS) technology provides an innovative experimental instrument for the research on vehicle emission controls. In addition, its advantages in terms of convenience, real-time nature, and massive data collection capability also provide a completely new approach to the development of vehicle driving cycles. The development of driving cycles is the basis for quantifying vehicle emissions. However, the existing driving cycles have many shortcomings such as insufficient base data, lack of basis in selecting parameters, fixed cycles without flexibility, lack of adequate evaluation methods, etc. Therefore, the research in this thesis is intended to achieve two objectives. The first objective is to develop driving cycles. In this regard, it develops driving cycles by utilizing the relationships between the characteristics of vehicle driving modes and emissions in combination with the principal component analysis in determining the primary influential parameters to the emissions. Then, it uses the cluster analysis to establish different driving cycles for the data with different driving modes and emission characteristics. The second objective is to evaluate driving cycles. In this regard, it adopts a parameter that can closely connect the driving modes and emissions, the Vehicle Specific Power (VSP), to evaluate whether the driving cycles can reflect the driving and emission characteristics on the real roads. It provides a reasonable explanation of the relationship between emissions and driving modes in the physical process of the driving, which is verified by a comparison of two parameters, speed and acceleration. Further, it develops a method to evaluate driving cycles by comparing the VSP bin distribution of driving cycles with the VSP bin distribution of the entire sampled data. Finally, it provides an example to show the feasibility of the proposed evaluation method.