A new mechanism for MPR selection in mobile ad hoc and sensor wireless networks

In mobile and sensor wireless networks, liaising between wireless nodes is performed using routing protocols. Each protocol has its advantages and drawbacks. OLSR is one of the most widely used protocols, due its proactive scheme. It is based on the principle of Multi Point Relay (MPR) which essentially consists in building a "relevant" set of links and direct neighbors by ignoring redundancies in order to have optimal paths. This relevance is demonstrated by criteria that allow the entire neighborhood of a node to be reached by two hops (all neighbors of neighbors). However, these criteria do not take into consideration other important factors that have a strong influence on the protocol's behavior such as the residual energy, which may lead to degrading the overall network lifetime. In this paper, we present a new method to overcome this problem, allowing mobile nodes within ad hoc and sensor wireless networks to determine an efficient set of their routers using OLSR Protocol, based on the energy constraint of their neighborhood. Due to the random mobility of nodes in such networks, an efficient calculation algorithm of routers for each node seems an obligation, in order to avoid link breakage and assure a high level of the network availability. Our new algorithm consists on introducing a dynamic weighting ratio between the reachability and the residual energy of the one hop neighbors of each node for selecting its routers. One of a statistical method to discover the neighborhood is the dispersion calculation of the residual energy that allows a node to find out how spreads are energies of its neighborhood. This calculation is introduced in the router selection algorithm. Thus, this method allows a load balancing in the network and avoids the rapid decrease of nodes batteries that have been selected as routers for a long time. Simulations results show that our new approach significantly increases the lifetime of the mobile ad hoc networks by decreasing the dead nodes number.