A5043313627- Autonomous Vehicle Technology and Safety
- Robotic Path Planning Algorithms
- Human-Automation Interaction and Safety
- Traffic Prediction and Management Techniques
- EEG and Brain-Computer Interfaces
- Robotics and Sensor-Based Localization
- Reinforcement Learning in Robotics
- Robot Manipulation and Learning
- Cognitive Science and Mapping
- Memory and Neural Mechanisms
- Robotics and Automated Systems
- Simulation Techniques and Applications
Hong Kong University of Science and Technology
2023-2024
University of Hong Kong
2023-2024
Beijing Institute of Technology
2021-2022
Abstract As intelligent vehicles usually have complex overtaking process, a safe and efficient automated system (AOS) is vital to avoid accidents caused by wrong operation of drivers. Existing AOSs rarely consider longitudinal reactions the overtaken vehicle (OV) during overtaking. This paper proposed novel AOS based on hierarchical reinforcement learning, where reaction given data-driven social preference estimation. incorporates two modules that can function in different phases. The first...
Developing efficient traffic models is essential for optimizing transportation systems, yet current approaches remain time-intensive and susceptible to human errors due their reliance on manual processes. Traditional workflows involve exhaustive literature reviews, formula optimization, iterative testing, leading inefficiencies in research. In response, we introduce the Traffic Research Agent (TR-Agent), an AI-driven system designed autonomously develop refine through iterative, closed-loop...
Transferring and reusing previously learned knowledge will enhance the performance of autonomous vehicles in newly encountered scenarios, hence is critical for fully driving. Previous data-level transfer methods fail to take into account scenario-level features belied those similar depend heavily on quality quantity data. In this paper, we provide a framework motion planning driving, named SceTL. By capitalizing successor representation, general among scenarios can be captured thereby...
Even driving in the same environment, different drivers may perform differently for encoding spatial information of which implies environmental cognition ability (ECA) among drivers. To assess such kind ECA, a brain-inspired computational grid cell model is used to mimic neuronal activation driver's brain when driving. Based on generated result, map, some analyses are conducted ECAs divers. A typical parking lot selected case study and data collected from simulator. The experimental results...