This content is not included in your SAE MOBILUS subscription, or you are not logged in.
Decision Tree Regression to Identify Representative Road Sections for Evaluating Performance of Connected and Automated Class 8 Tractors
Technical Paper
2021-01-0187
ISSN: 0148-7191, e-ISSN: 2688-3627
Sector:
Event:
SAE WCX Digital Summit
Language:
English
Abstract
Currently, connected and autonomous vehicle (CAV) technology is being developed for class 8 tractor trucks aimed at improved safety, fuel economy, and reduced CO2 emissions. Despite extensive efforts that have been conducted across the world, the reported efficiency gains were varied from different research groups which raised concerns about the fidelity of models, the performance of control, and the effectiveness of the experimental validation. One root cause for this variation stems from the fact that the efficiency gain obtained from the CAV is sensitive to real-world conditions, including surrounding traffic and road grade. This study presents an approach aimed to identify representative public road section and facilitate the CAV research from this perspective. By employing the decision tree regression method to the Fleet DNA database, the most representative road sections can be identified. High-level metrics and detailed information of the derived road sections are also illustrated and discussed, which demonstrate their representativeness and the effectiveness of the approach. Meanwhile, the capability of this approach can be easily extended by integrating specific constraints into the DTR algorithm. As an example, a specific representative road section with an aggressive road grade profile was also provided via this approach.