Joint congestion and medium access control in cooperative vehicular ad-hoc networks

In this paper, we address the congestion and medium access control (MAC) problem for cooperative vehicular ad-hoc networks (VANETs) in the framework of utility function maximization. Compared to its counterpart in a wired network where link capacities are assumed to be fixed, congestion control in a multi-hop cooperative VANETs is much more complex and requires joint optimization at both the transport layer and the MAC layer. This is due to the fact that the attainable function value on each path in the network is "elastic" which results in typically non-convex (non-coucave) or non-separable function of the transmission attempt rates. A new cross-layer algorithm is proposed to solve the system problem with constraints that arise from the cooperation among the users and contention for the wireless channel access. The algorithm works at the MAC layer to adjust link persistence probability and at the transport layer to adjust flow rates joitly, and it is designed in distributed manner that is amenable to on-line implementation for VANETs. We prove rigorously that the proposed algorithm converges to the globally optimal solution.