Simulation and verification of autonomous route planning behavior

[摘要] This thesis presents a new dynamic traffic simulator called DUST, used to investigate new foundational theory for autonomously reconfigurable cyber-physical systems in the presence of unexpected disruptions. We focus on transportation networks to develop our methodologies since it is a prime example of a cyber-physical system that is characterized by distributed decision-making and is prone to unexpected disruptions. This new simulator investigates the route choice behavior of cars in an urban environment. It allows subjects to participate as drivers in virtual city environments directly as well as artificial intelligence algorithms. It provides a platform for experimentation with various incentive mechanisms and information dissemination protocols that are critical for emergency planning during disruptions.

[发布机构] Massachusetts Institute of Technology