Automation Of Road Intersections Using Consensus-based Auction Algorithms

TL;DR

This paper explores a decentralized traffic control method for automated road intersections using consensus-based auction algorithms, where vehicles negotiate crossing order and compute collision-free trajectories with onboard MPC, demonstrated through simulations.

Contribution

It introduces a novel decentralized intersection control approach combining auction algorithms with model predictive control for collision avoidance.

Findings

Effective negotiation of crossing order among vehicles.

Successful simulation of collision-free trajectories.

Potential for improved traffic flow and safety.

Abstract

This paper investigates a consensus-based auction algorithm in the context of decentralized traffic control. In particular, we study the automation of a road intersection, where a set of vehicles is required to cross without collisions. The crossing order will be negotiated in a decentralized fashion. An on-board model predictive controller (MPC) will compute an optimal trajectory which avoids collisions with higher priority vehicles, thus retaining convex safety constraints. Simulations are then performed in a time-variant traffic environment.