University of Twente Student Theses
As of Friday, 8 August 2025, the current Student Theses repository is no longer available for thesis uploads. A new Student Theses repository will be available starting Friday, 15 August 2025.
Driving For Endurance : Fuel and Tire Efficient Autonomous Racing in Assetto Corsa Using Reinforcement Learning
Rusk, Nathan (2025) Driving For Endurance : Fuel and Tire Efficient Autonomous Racing in Assetto Corsa Using Reinforcement Learning.
|
PDF
11MB |
| Abstract: | Reinforcement learning (RL) has shown promise in training autonomous racing agents, yet most current approaches optimize solely for lap time, neglecting critical endurance factors like fuel consumption and tire wear. This paper explores how RL agents can be trained to balance competitive racing performance with long-term efficiency in the high-fidelity racing simulator Assetto Corsa. Building on state-of-the-art Soft-Actor-Critic (SAC) based frameworks, we extend both the observation space and reward function to include endurance metrics. Several agents are trained and evaluated on their ability to manage fuel and tire resources while maintaining speed across multi-lap sessions. Quantitative and observational analyses show that these methods play a crucial role in enabling efficient driving behaviors. A balanced agent improved fuel and tire wear efficiency by 22.6% and 10.5%, with realistic racing strategies such as lift and coast emerging spontaneously. However, trade-offs in lap time remain a challenge. These findings contribute to the growing field of multi-objective reinforcement learning in simulated environments and lay the groundwork for more realistic and strategic autonomous racing agents. |
| Item Type: | Essay (Bachelor) |
| Faculty: | EEMCS: Electrical Engineering, Mathematics and Computer Science |
| Subject: | 54 computer science |
| Programme: | Business & IT BSc (56066) |
| Link to this item: | https://purl.utwente.nl/essays/107596 |
| Export this item as: | BibTeX EndNote HTML Citation Reference Manager |
Show download statistics for this publication
Show download statistics for this publicationRepository Staff Only: item control page