Adaptive Clustering Protocol for Wireless Networks

We present a new protocol to cluster nodes of wireless networks. Our protocol ensures reliable message transmission in spite of collisions and transmission losses. We first present a geometric-probabilistic model to predict the optimal range for maximizing one-hop broadcast coverage in wireless networks as a function of range, sending rate and density. We use the above model to design a scalable clustering protocol, which functionality is independent of network topology. Then, we incorporate our analytical model to this protocol to adapts to network conditions. Through extensive analytical and experimental results, we show multifold advantages of adapting the protocol to network conditions.