An Energy Efficient Distributed Gossip Algorithm for Wireless Sensor Networks based on a Randomized Markovian Duty-Cycling

TL;DR

This paper introduces an energy-efficient distributed gossip algorithm for wireless sensor networks that combines asynchronous consensus with a Markovian duty-cycling mechanism, ensuring convergence and stability.

Contribution

It presents a novel asynchronous consensus algorithm with a Markovian duty cycle, enhancing energy efficiency while guaranteeing convergence in wireless sensor networks.

Findings

Algorithm converges to the network average consensus.

Demonstrates stability across various graph topologies.

Energy efficiency is improved through Markovian duty-cycling.

Abstract

This paper proposes a novel asynchronous consensus algorithm which is based on a continuous update rule and an energy efficient event triggered duty (wake-sleep) cycle based on a discrete Markov chain model. The system model of the proposed algorithm is formulated and analyzed. The conditions for convergence and stability of the algorithm are derived and the algorithm is proved to converge to an average consensus. Numerical simulations on random, circular, chain and star graphs show stability in consensus and convergence to network global state average.