A Decentralized Approach to Minimum-Energy Broadcasting in Static Ad Hoc Networks

Broadcasting is a commonly used feature in wireless networking, e.g. for file distribution, re-tasking, event notification, or miscellaneous maintenance. Due to the limited resources of most wireless ad hoc and sensor networks, minimizing the cost of broadcasts is of utmost importance. This has led to work in the minimum energy broadcasting problem. Solutions to this issue have primarily focused on topology control algorithms or broadcast oriented protocols. Many of these solutions are based on a centralized protocol, where a node has global topology knowledge of the network. However, this information is typically not available in ad hoc applications. Decentralized approaches to this problem have not been able to match the energy efficiency of centralized methods. Both centralized and decentralized approaches have relied upon locality information to estimate link cost. This method is unreliable and can result in poor network models. In this paper, we will describe a new distributed approach to the minimum energy broadcasting problem. Our protocol constructs a broadcast tree in a distributed fashion using only local information. It uses link quality measurements to more accurately estimate link cost, which makes it adaptable to any environment. We show by simulation that our protocol is capable of constructing a tree that is near to centralized approaches in energy cost, without requiring an excessive overhead cost.