Sim2Real Transfer for Traffic Signal Control

Traffic signal control is a complex and important task that affects the daily lives of millions of people. Reinforcement Learning (RL) has shown promising results in optimizing traffic signal control, but transferring learned policies from simulation to the real world remains a challenge due to the domain gap between the simulation and the complex real-life scenario. In this paper, we utilize grounded action transformation to mitigate the domain shifting problem and improve Sim2Real transfer for RL-based traffic signal control. Grounded action transformation leverages the dynamics between the simulation and real-world actions to generate effective real-world actions. We evaluate our method on a simulated traffic environment and show that it significantly improves the performance of the transferred RL policy in the real world. Our results demonstrate the potential of grounded action transformation as a promising technique for Sim2Real transfer in RL-based traffic signal control.