Robust Lightless Intersection Control Framework for Heterogeneous Vehicles

The growing popularity of autonomous vehicles and their integration into everyday life has shifted the focus of their development towards ensuring the safety of these self-driving entities, commonly referred to as ego vehicles. A significant breakthrough in this domain is the ability of interconnected vehicles to seamlessly navigate intersections in the absence of traffic lights. This paper presents a novel robust control architecture for heterogeneous autonomous vehicles, tailored to uphold safety and efficiency in such light-less intersections, particularly in the face of disruptions. To accommodate heterogeneous vehicles, the proposed architecture is structured around two core layers: the consensus layer and the target-following layer. The vehicles communicate only through the consensus layer and do not share any physical information. In the consensus layer, autonomous vehicles engage in a collaborative process, sharing each other's virtual times. These virtual times represent the scheduled moments for the vehicles to be at specific points on their predetermined paths. Upon establishing the virtual time and path, each vehicle is allocated a virtual target. The target-following layer plays a crucial role in guiding each vehicle to adhere to its virtual target. In instances where a vehicle is impeded by a disturbance and consequently lags, it recalibrates its virtual time and communicates this revised timing to the rest of the vehicles via the consensus layer. This collective adjustment results in a moderated pace for all vehicles in the network, thereby bolstering safety and robustness while maintaining overall efficiency. Furthermore, the paper delves into the time shift allocation of virtual times for each path, ensuring collision-free operation. The paper meticulously outlines the design of this architecture and validates its efficacy in ensuring the safe and synchronized operation of autonomous vehicles at intersections without traffic lights. This paper details the design and validation of the architecture, demonstrates its effectiveness through simulation, and discusses the implications of transitioning to fully autonomous intersection management.