WISE: Efficient Wireless Access Point SElection for Large-Scale Cooperative System

Cooperative design is becoming especially import for loosely coupled MIMO system. Cooperative AP selection is one of the most important issue need to be addressed. In this study, we investigate the efficient AP selection for cooperative MIMO system. The major concern in this work is to reduce the coordinative overhead in network. Two key challenges need to be formally addressed before achieving aforementioned goal. First, the decoding symbols in network is moderately large, which will lead to considerably high network traffic. Second, the decoding performances are highly correlated with the decoding symbols. Reducing symbol transfer will inevitably lead to low decoding rate. Thus, high quality decoding data, say, selecting the appropriate AP for decoding and performing efficient symbol transfer would be a possible solution to tackle aforementioned difficulties. In our design, we fully evaluate the performance of symbol level decoding for cooperative MIMO system, and explore the relationship between the cooperative AP selection with overall decoding. We propose WISE: a simple but effective scheme for AP selection, which is almost optimal and aggressive with moderately low complexity. Our algorithm is implemented and evaluated with USRP, an open software radio network design platform, which enable us to evaluate and validate our scheme more effectively. Experimental results show that, ours could significantly reduce the network overhead, and maintain relatively high decoding rate for cooperative MIMO system.