Optimal deployment of impromptu wireless sensor networks

Abstract

The need for impromptu wireless networks arises in emergency situations where the team responding to the emergency, needs to deploy sensors (such as motion sensors, or even imaging sensors) and a wireless interconnection network, without any prior planning or knowledge of the terrain. In this paper, we consider a simple model for the sequential deployment of wireless relays as a person steps along a “corridor” of unknown length, so as to create a multihop network for interconnecting a sensor to be placed at the end of the corridor with a control truck standing near the entry to the corridor. Assuming low traffic and simple link-by-link scheduling, we consider the problem of minimising an end-to-end cost metric (e.g., delay or power from the sensor to the control centre) subject to a constraint on the number of relays. Two kinds of constraints are considered: the expected number of relays is bounded, or the actual number of relays is bounded. In each case, the problem is formulated as a Markov decision process. The problem of deciding whether or not to place a relay at each step is shown to be equivalent to a certain stochastic shortest path problem embedded at relay placement points. Numerical results are provided to illustrate the performance trade-offs.