The overall diagram of this study (IMAGE)
Caption
It contains two main parts: fNIRS risk detection and fNIRS integrated reinforcement learning (RL). The first part covers real-time pre-processing, feature extraction, and classification models. The second part focuses on the proposed human-guided deep reinforcement learning (DRL) scheme, and it contains a TD3 agent with some modifications, an intelligent driver model (IDM), and a human-guided DRL switching mechanism. This switching mechanism may accelerate the learning speed of TD3 based on passengers’risk assessment using fNIRS.
Credit
Xiaofei Zhang, School of Vehicle and Mobility, Tsinghua University.
Usage Restrictions
News organizations may use or redistribute this image, with proper attribution, as part of news coverage of this paper only.
License
Licensed content