Skip to Main Content
In this paper we consider improving the quantity and quality of data collected and sent to a sink over the lifetime of a network of mobile sensors. The positions of the sensors and the routing strategy chosen are the variables in our problem. The original problem is a multi-objective non-convex optimization problem and we believe it to be tough to solve. The approach taken in this paper is to break down the original problem into sub-problems and develop an iterative scheme to optimize both the quantities. We propose three sub-problems. First, we optimize the way (obtain flows) the data is sent to the sink for a fixed placement of the nodes by solving a linear program (work on this has already been done in the past by others). Once flows have been obtained, they are kept constant and the nodes are then moved in a way such that the lifetime is further improved, keeping the maximum distortion error less than or equal to what it is for the initial node placement. Then for the new node distribution with a better lifetime, we decrease the maximum distortion error keeping the lifetime greater than or equal as compared to the starting configuration. Centralized and decentralized iterative schemes are presented using these subproblems that monotonically improve the lifetime data gathering and reduce maximum distortion error at each step.
Date of Conference: 4-7 Oct. 2004