Distributed Algorithm for Dynamic Data-Gathering in Sensor Network

TL;DR

This paper introduces a distributed, energy-efficient data-gathering algorithm for sensor networks that adaptively updates data paths to extend network lifetime without requiring global topology knowledge.

Contribution

The proposed algorithm computes and periodically updates energy-efficient data paths in a distributed manner, significantly improving sensor network lifetime over existing methods.

Findings

Algorithm achieves higher network lifetime than MST, SPT, and WRT schedules.

Schedule updates are efficient, using minimal energy and only local information.

The algorithm adapts to residual energy, prolonging system operation.

Abstract

In WSN, each sensor is responsible for sensing environmental conditions and sending them to the one or more base stations. Battery-operated sensors are severely constrained by the amount of energy that can be spend for transmitting these sensed data. However, aggregation of data (including removal of redundant data) at intermediate sensors and forwarding of aggregate data reduce overall energy consumptions in WSN. In general, data gathering refers to the process of periodic collection of sensed data from various sensors to one or more base stations (BS). Energy efficient data gathering scheduling is essential for improving the lifetime of WSN. In this paper, we propose a distributed algorithm to compute data-gathering schedule that aim to improve the lifetime of WSN by suitably selecting energy-efficient data-flow paths from various sensors to the base station. For a multihop WSN with $n$ sensors, the proposed algorithm first computes a schedule in $O(n^2)$ time steps, and then this schedule is periodically updated based the residual energy and the feedback received from the BS. The system performs approximately $\log(\mathcal{L})$ schedule updates where $\mathcal{L}$ is the expected lifetime of the system in number of data-gathering rounds. Moreover, each updation process uses the existing active schedule (data-flow path) - thus consuming only a small fraction of a single data gathering round activity. Such an algorithm thus could precisely incorporate the energy consumptions due to updates and related activities. Moreover, our algorithm does not assume any global knowledge of the topology or the positions of various sensors. Through simulation study, we found that our proposed algorithm achieves significantly higher network lifetime compared to existing data-flow schedules based on the Minimum Spanning Tree (MST), the Shortest Path Tree (SPT), the Weighted Rooted Tree (WRT).