Abstract

This research addresses the critical challenge of path planning for autonomous vehicles by integrating Social Value Orientation (SVO) into path-planning algorithms for autonomous vehicles. The framework utilizes a fuzzy logic-based system to evaluate and categorize the social values of pedestrians and vehicles in real time by considering their observed behaviors and social cues. This approach enables autonomous vehicles to make more informed and socially aware decisions, thereby enhancing their ability to interact safely with human road users. In addition, the thesis introduces an adaptive artificial potential field (APF) method that dynamically assesses the dangers presented by different road users, taking into account factors such as type, size, speed, and societal importance. By integrating road layout and traffic signal potential fields, the APF method guarantees that the autonomous vehicle can navigate intricate road conditions while prioritizing safety. The effectiveness of the proposed path planning framework is rigorously validated using CARLA driving simulator. These simulations create a realistic and dynamic traffic environment, allowing for the thorough testing of custom behavioral profiles for various actors in different traffic situations. Results demonstrate the framework's practical applicability and effectiveness in enhancing the interaction between autonomous vehicles and human road users.
The outcomes of this research contribute to the development of safe and human-centric path-planning algorithms for highly automated vehicles, particularly in dense or mixed-traffic environments. This work represents a significant step towards creating autonomous systems that can coexist harmoniously with human drivers and pedestrians, ultimately leading to safer and more efficient roadways.