Trajectory Design of Connected and Automated Vehicles with Pattern Recognition of Surrounding Human-Driven Vehicles (CAVTD)

Research Project, 2019

Connected and Automated vehicle (CAV) technique shows a promising future of traffic with significantly improved efficiency, stability, and safety. In the next few decades, however, it is inevitable that CAVs have to drive among human drivers in a mixed environment. This leads to a challenge that CAVs need to accommodate the inherently stochastic nature of human-driven traffic flows, to perform reasonably. The current fashion of addressing this challenge is to obtain the location of surrounding vehicles with V2V networks or high-frequency on-board sensors and adjust CAV trajectories accordingly. To this end, the present project proposes a memory-based trajectory design method that could improve the performance of current methods.