Cooperative Driving:

We tackle the problem of autonomous driving in complex & competitive mixed-autonomy environments where autonomous vehicles interact with vehicles driven by humans. Formally, we model this problem as a partially-observable stochastic game and train reinforcement learning agents that cooperate with each other and sympathize with human-driven vehicles. Our autonomous agents learn general latent representations that enables them to coordinate in novel environments. For more details about each paper, please refer to the paper’s individual website:

Publications:

Stanford-Logo
ucsb

Cooperative Autonomous Vehicles that Sympathize with Human Drivers [submitted]

Behrad Toghi, Rodolfo Valiente, Dorsa Sadigh, Ramtin Pedarsani, Yaser P. Fallah

2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2021)

cooperative driving
Stanford-Logo
ucsb

Altruistic Maneuver Planning for Cooperative Autonomous Vehicles Using Multi-agent Advantage Actor-Critic

Behrad Toghi, Rodolfo Valiente, Dorsa Sadigh, Ramtin Pedarsani, Yaser P. Fallah

2021 ADP3 Workshop at IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR 2021)

cvpr
Stanford-Logo
ucsb

Social Coordination and Altruism in Cooperative Autonomous Driving [submitted]

Behrad Toghi, Rodolfo Valiente, Dorsa Sadigh, Ramtin Pedarsani, Yaser P. Fallah

IEEE Transactions on Intelligent Transportation Systems (T-ITS)

tiv

Behrad Toghi Behrad Toghi Behrad Toghi Behrad Toghi