Safe and efficient interaction with the environment requires an autonomous system to understand the effects of its actions on the decisions of other agents in the scene. State-of-the-art model-based optimal control approaches treat other agents as dynamic obstacles and ignore their rational behavior. In contrast, game-theoretic planning allows to model differing objectives of multiple agents explicitly. The goal of this project is to implement both techniques and compare their performance in a realistic simulated environment.
- Lasse Peters (mentor)
- Dhagash Desai
- Saurabh Gupta