Data-Driven Synthetic Optimization Method for Driving Cycle Development

• Main Authors: Renjuan Sun, Yuxin Tian, Hongbo Zhang, Rui Yue, Bin Lv, Jingrong Chen
• Format: Article
• Language: English
• Published: IEEE 2019-01-01
• Series: IEEE Access
• Subjects: Synthetic optimization; driving cycle; data-driven
• Online Access: https://ieeexplore.ieee.org/document/8886437/

One of the key tasks for improving vehicle operating costs estimation is to develop representative driving cycles. A driving cycle is a vehicle speed-time profile. The cycles are a critical input for simulating VOCs in different road scenarios. The traditional methods could not generate driving cycles representing the speed pattern of the sample snippets. A new driving cycle development method, synthetic optimization, was developed and applied to generate synthetic driving cycles by applying the speed-acceleration frequency distribution matrix, speed-acceleration status transition matrix, and simulated annealing optimization algorithm. The data used for driving cycle development come from the new Strategic Highway Research Program (SHRP 2) naturalistic driving study (NDS) data and truck GPS trajectory data from American Transportation Research Institute (ATRI). Driving cycles of full-access-control facilities were developed as an example to show the performance of the proposed method. Compared to the conventional methods, the synthetic optimization approach provides, along with other advantages, driving cycles that better represent the speed patterns observed in the different scenarios.