This research studies new methods for constructing economical plans to deploy wireless sensor nodes of multiple types and form a reliable wireless sensor network to monitor, throughout a given time period, a large set of targets spread across a geographical space. It initiates and carries out a thorough analytical investigation on deploying multiple types of sensors for constructing a wireless sensor network that satisfies, simultaneously, the min-cost, the lifetime, the fault-tolerance, and the connectivity requirements. It integrates concepts and methods in linear programming, convex programming, reliability theory, graph theory, and approximation algorithms to model these requirements and find efficient solutions. Results of this research provide a substantial impact on large-scale, terrestrial and underwater, wireless sensor network applications and communication protocol designs.
More specifically, this research seeks analytical models to provide, with minimum monetary cost on sensor nodes, reliable sensor placements and time scheduling to hinge on spatial-temporal information and power consumption. Despite failures and limited power supply of sensor nodes, the network must guarantee that each spatial target be watched by at least one active sensor at any given moment during a pre-determined period of time required by the underlying application. This implies sensor placements must satisfy the spatial-temporal coverage, the fault-tolerance, and the min-cost requirements for single-hop networks; as well as the additional connectivity requirement for multi-hop networks, which ensures that each sensor node is connected to a base station through a communication path.
