Traffic diffusion analysis for adaptive multi-path routing algorithm in sensor networks

Abstract

Multipath routing algorithms in sensor networks provide extensive scope for adaptively routing and dispersing traffic over the network. An essential requirement for adaptive routing in sensor networks is to account for the high cost associated with the use of battery power and network bandwidth. In this paper, we propose an adaptive multi-path routing algorithm (AMRA) that diffuses the traffic such that energy consumption in minimized. We define a Hamiltonian as a function of the cost incurred and the benefit obtained by a node for forwarding packets. The Hamiltonian is then subjected to Pontryagin's cost minimization for efficiently diffusing the packets among the next-hop nodes. This traffic diffusion is computed by solving the steady state queue length equation which we represent in the frequency domain, ns-2 based simulation experiments are conducted and the performance of the proposed algorithm is compared against the AODV routing protocol.