Efficient algorithms for MAC layer duty cycling and frame delivery in wireless sensor network

Thomson, Craig James

Abstract

In Wireless Sensor Networks, with small, limited capacity devices now more prevalent, the issue of Neighbour Discovery has shifted. These devices utilise duty cycling methods in order to conserve battery power. Hence, the main issue is now that these devices may be awake at the same time in order to discover each other. When mobility increases complexity further. Rather than attempt to negate the issue of mobility, instead this thesis seeks to utilise a predictable sink mobility pattern in order to influence the duty cycling of static nodes.

Literature demonstrates a move towards Mobility Awareness in Neighbour Discovery in mobile Wireless Sensor Networks. However, there is a gap identified with sink mobility in use. Therefore, this thesis aims to establish to what extent the mobility pattern of a Mobile Sink Node in a Wireless Sensor Network may be exploited at the MAC layer, to influence the performance of static nodes. Such that network efficiency may be improved with energy consumption reduced and balanced across nodes. This study proposes three novel lightweight algorithms, with processing which does not add to the energy consumption within sensor nodes, these being Mobility Aware Duty Cycling Algorithm (MADCAL), Mobility Aware Duty Cycling and Dynamic Preambling Algorithm (MADCaDPAL) and Dynamic Mobility and Energy Aware Algorithm (DMEAAL). These located in the MAC layer of static nodes and utilising knowledge of predictable sink node mobility. This is in order to create a dynamic communication threshold between static nodes on the sink path and the sink itself. Subsequently lessening competition for sink communication between nodes.

In MADCAL this threshold is used to influence the sleep function in order that static nodes only awake and move to Clear Channel Assessment once the sink is within their threshold, improving energy consumption by up to 15%. The MADCaDPAL algorithm takes this approach further, using the threshold to directly influence Clear Channel Assessment and the sending of preambles, as such, closing off the threshold when the sink leaves it. This shows energy consumption lessening by close to 80% with a significant improvement in frame delivery to the sink. Finally, the DMEAAL algorithm utilises previous results to influence energy consumption in real-time by utilising a cross-layer approach, comparing current consumption to optimal target energy consumption and adjusting the threshold for each static node accordingly. This shows benefit in evening out results across nodes, thus improving network lifetime. All algorithms are achieved without the energy-consuming beacon messaging associated with Neighbour Discovery. Analysis and simulation results, tested on a lightweight implementation of a carrier-sense multiple-access-based MAC protocol, show a significant improvement in energy consumption and frame delivery in both controlled and random environments. In utilising a cross-layer approach to access energy consumption in static nodes, is it also shown to be possible to even out energy consumption across nodes by altering the communication threshold in real-time. As such, improving network lifetime by removing spikes in energy consumption in individual nodes.

Citation

Thomson, C. J. Efficient algorithms for MAC layer duty cycling and frame delivery in wireless sensor network. (Thesis). Edinburgh Napier University. Retrieved from http://researchrepository.napier.ac.uk/Output/2850793

• Thesis Type: Thesis
• Deposit Date: Mar 4, 2022
• Publicly Available Date: Mar 4, 2022
• DOI: https://doi.org/10.17869/enu.2022.2850793
• Public URL: http://researchrepository.napier.ac.uk/Output/2850793
• Award Date: Aug 31, 2021