Adaptive Network Coding for Wireless Access Networks

We propose a framework for optimizing the quality of service of multiple simultaneous flows in wireless access networks via network coding. Specifically, we consider the typical scenario in which multiple flows originate from multiple sources in the Internet and terminate at multiple users in a wireless network. In the current infrastructure, the wireless base station is responsible for relaying the packets from the Internet to the wireless users without any modification to the packet content. On the other hand, in the proposed approach, the wireless base station is allowed to perform network coding by appropriate linear mixing and channel coding of packets from different incoming flows before broadcasting a single flow of mixed or coded packets to all wireless users. Each user then uses an appropriate decoding method to recover its own packets from the set of coded packets that it receives. We show that in principle, for the given channel conditions and QoS requirements, appropriate mixing and channel coding of packets across different flows can lead to substantial quality improvement for both real-time and non-real time flows. On the other hand, blind mixing can be detrimental. We formulate this mixing problem as a combinatorial optimization problem, and propose a heuristic algorithm based on simulated-annealing method to approximate the optimal solution. Simulation results verify the performance improvement resulting from the proposed approach over the non-network coding and the state-of-the-art network coding approaches.