A Feasible Human-like Multidimensional Driving Behavior Framework Design for Autonomous Vehicles

Artificial intelligence (AI) technology is being used to drive vehicles autonomously, which helps to enhance driving safety by eliminating human error. However, the AI driver might make unreasonable decisions, which reduces its trustworthiness to humans because driving is a complex problem based on nonverbal communication among drivers. To elevate driving trustworthiness, understanding human characteristics using a multiagent driving simulator and modeling them using AI algorithms are key components. This dissertation aims to design a feasible human-like driving behavior framework to minimize the inconvenience of driving. This dissertation will contribute to the development of human-like driving behavior for autonomous vehicles.