F1tenth is an international community performing autonomous systems research through building and racing of 1:10 scale F1 vehicles. The common goal is to race around tracks as fast as possible – usually first in simulation and then in the real world. This provides a platform for addressing many research and systems engineering questions and solving both theoretical and practical problems. The current approaches to autonomous racing ranges from more conventional control approaches to data-driven solutions involving machine learning.
Since the dynamics and sensor models are often imperfect and potentially even unavailable in these settings, learning from data will very likely be part of most optimal solutions to autonomous racing. This is further supported by the large amounts of data and fast computers available with modern technology. One popular approach is to use system identification to estimate these models, followed by an optimization method such as model predictive control to calculate the optimal trajectory. A separate but related alternative from the field of computer science is to use a model-based variant of reinforcement learning, where the idea is to supplement the learning from actual experience with simulated experience. The aim of this project is to compare these two approaches for the task of overtaking in multi-vehicle racing on a physical track.