We found a match
Your institution may have rights to this item. Sign in to continue.
- Title
Advanced Sensing and Safety Control for Connected and Automated Vehicles.
- Authors
Huang, Chao; Wang, Yafei; Hang, Peng; Zuo, Zhiqiang; Leng, Bo
- Abstract
Advanced sensing technologies and control algorithms, working to acquire environmental data, analyze data, and regulate vehicle movements, are key functional components of CAVs. The proposed SGRL algorithm uses the training approach for a single agent, constructs a clearer incentive reward function, and greatly increases the dimension of the action space over the conventional deep neural network (DQN) algorithm. Ref. [[7]] proposes a generalized single-vehicle-based graph neural network reinforcement learning (SGRL) algorithm that incorporates interactive information between agents in the environment into the decision-making process of autonomous driving. The experimental results on an automated patrol vehicle indicated that the single-sensor algorithm's pedestrian detection results could be improved by more than 17% and that the multi-layer fusion method provided more understandable density estimation results.
- Subjects
AUTONOMOUS vehicles; REINFORCEMENT learning; OBJECT recognition (Computer vision); EUCLIDEAN algorithm; VISUAL odometry; INTERPOLATION spaces
- Publication
Sensors (14248220), 2023, Vol 23, Issue 2, p1037
- ISSN
1424-8220
- Publication type
Article
- DOI
10.3390/s23021037