Energy-Aware Path Following: Comparative Analysis of Reinforcement Learning and NMPC for Electric Vehicles
Researchers compare PPO and NMPC for energy-aware EV path following, with PPO transferring to unseen tracks.
The paper compares NMPC, PPO, a gain-scheduled Ackermann state-feedback baseline, and a Stanley geometric controller for energy-aware electric-vehicle path following. All share a Frenet kinematic model and the VT-CPEM energy model with regenerative braking, and NMPC is implemented with JIT-compiled CasADi. PPO trained on straight and S-curve tracks transfers to ISO 3888-1 lane changes, chicanes, random splines, and graded roads, and zero-shot to a dynamic single-track model before brief fine-tuning.
- Compares NMPC, PPO, PID-SF, and Stanley under one energy model
- PPO transfers unmodified to unseen tracks including ISO 3888-1
- Policy zero-shot transfers to a dynamic single-track model
- VT-CPEM energy model includes explicit regenerative braking
From the original article
1 min read at arxiv.orgPath-following control strategies typically follow the bi-objective optimization dilemma: minimizing deviations from a reference path while maintaining smooth speed profiles. The latter objective is especially relevant for Electric Vehicles (EVs), since their limited driving range can be extended by recovering energy through regenerative braking, a feature that has not yet been sufficiently studied in the literature.
By Mohamed Sabaa · arxiv.org
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.