Learning Realistic Traffic Agents in Closed-loop
The paper presents a method called Reinforcing Traffic Rules (RTR) for simulating realistic traffic scenarios, which is essential for the development of self-driving software. Traditional imitation learning methods often result…
Continue reading