A significant barrier to deploying autonomous vehicles (AVs) on a massive scale is safety assurance. Several technical challenges arise from the uncertain environment in which AVs operate, such as road and weather conditions, the uncertain behavior of pedestrians and agent vehicles, as well as model inaccuracy. In this project work, we will develop an algorithm pipeline along with a technical stack for trajectory optimization for AVs with bounded risk of collision. We consider three major contributions. First, an intention recognition system that predicts the driving-style and the intention of agent vehicles and pedestrians. Second, a planning system that takes into account the uncertainty in the environment and propagates all the way to control policies that explicitly bound the risk of collision. Third, a high-level planning system that extends to multi-vehicle on-the-spot collaboration, under diverse game-theoretic situations that arise due to competing local objectives among AVs.