NPC: Neural Predictive Control for Fuel-Efficient Autonomous Trucks

Jiaping Ren, Jiahao Xiang, Hongfei Gao, Jinchuan Zhang, Yiming Ren, Yuexin Ma, Yi Wu, Ruigang Yang, Wei Li·December 18, 2024

Summary

A Neural Predictive Control (NPC) method optimizes fuel efficiency for autonomous trucks, using data-driven models to predict vehicle dynamics, road slope, and fuel consumption. This approach outperforms traditional Predictive Cruise Control (PCC), saving 2.41% and 3.45% more fuel in simulations and open-road testing, respectively. The NPC framework, incorporating an attention-based NVFormer, enhances fuel-saving optimization by predicting missing future data components. This method is robust across varying conditions, offering a promising solution for improving fuel consumption efficiency in autonomous truck operations.

Key findings

3

Advanced features