Analytical revisit on the use of biased random walks for data forwarding in wireless sensor networks

Abstract

In recent years, the use of random walks (RW) for data forwarding in wireless sensor networks (WSN) has gained a lot of popularity. However, a negative effect is often caused by the fact that a completely uniform random choice of the next hop during the walk translates into a reduced progress towards the sink node, thereby yielding a long latency. Several strategies are proposed to attenuate this problem. Most of them are characterized by their dependence on state information stored in sensor nodes in order to bias the direction of the walk towards the target. Such information require additional capabilities and it is costly in terms of complexity and energy, which is inherently problematical in WSN. This leads to a tradeoff between mutually contradictory goals and raises the question as to what extent biasing RW can affect the performance of the data forwarding scheme. This question is our primary motivation in investigating the problem of biasing RW based data forwarding from a pure analytical perspective.