Uncertainty-aware Connected Autonomous Vehicles Stack: Perception, Intention Recognition, Planning and Control.

A significant barrier to deploying connected autonomous vehicles (CAVs) on a massive scale is safety assurance. Several technical challenges arise from the uncertain environment in which autonomous vehicles (AVs) operate, such as road and weather conditions, errors in perception and sensory data, Vehicle-to-Vehicle (V2V) communication loss, and the uncertain behavior of pedestrians and agent vehicles. In this proposed work, we will develop an algorithm pipeline and a technical stack for probabilistic scene awareness and trajectory optimization with a bounded risk of collision.