Fair Transmission Rate Adjustment in Cooperative Vehicle Safety Systems Based on Multi-Agent Model Predictive Control

Abstract

Cooperative vehicle safety systems (CVSSs) rely on vehicular networking for broadcasting state information in order to track neighbors’ positions and, therefore, to predict potential collisions. In vehicular networking, a large number of vehicles compete for access to the limited channel resource, causing channel congestion. The vehicle tracking accuracy, which is the basis for CVSSs, therefore, can be heavily affected. Moreover, the available channel resources must be shared among vehicles in a fair way in order to maintain accurate tracking accuracy for each vehicle. To realize fair access to channel resources while maintaining an accurate tracking performance under conditions of dynamic vehicle density, in this paper, we present a distributed fair transmission rate control strategy, based on multi-agent model predictive control (MPC). We first propose a dynamic information dissemination rate model to capture the state information dissemination ability under conditions of dynamic vehicle density. Then, we present a multi-agent information dissemination model, in which each vehicle is controlled by a control agent that uses MPC and coordinates with its neighboring agents in order to determine its optimal transmission rate actions. We then design an augmented-Lagrangian-based distributed decision-making scheme to find the optimal transmission rate actions and, at the same time, reach an agreement on fair and efficient channel utilization among the vehicles. Simulation results confirm that the distributed transmission rate control strategy can guarantee fair access to channel resources while achieving the optimal vehicle tracking performance under conditions of dynamic vehicle density.