An MPC-based Task Priority Management Approach for Connected and Automated Vehicles Reference Tracking with Obstacle Avoidance

We present a reference tracking control problem with obstacle avoidance for connected and automated vehicles. The proposed approach allows to deal with obstacles in the control loop in the first instance, coping with real-time operations while safely waiting for an eventually new planned trajectory from the guidance loop. An obstacle avoidance algorithm is designed to produce suitable constraints for an optimization control problem to be solved via Nonlinear Model Predictive Control. Such an algorithm is based on task priority management, so that the reference tracking task is handled as a lower priority task with respect to the obstacle avoidance task. Automated vehicles are managed in a decentralized fashion, so that they can process independently any sensed potential obstacles, including conventional vehicles. In the presence of vehicle connectivity, vehicles may exchange information about their states to make decisions based on more accurate predictions. The proposed method is evaluated via simulation experiments, for a set of scenarios in the context of urban traffic.