Contextual dishonest behaviour detection for cognitive adaptive charging in dynamic smart micro-grids

Radenkovic, Milena and Walker, Adam (2018) Contextual dishonest behaviour detection for cognitive adaptive charging in dynamic smart micro-grids. In: 15th Wireless On-demand Network systems and Services Conference (WONS2019), 22-24 January 2019, Wengen, Switzerland. (In Press)

[img]
Preview
PDF - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader
Download (481kB) | Preview
[img] Plain Text (ZIP compressed plain text archive, containing: Data: Comma Separated Variables [.csv], Text [.txt] and Well-known text [.wkt] files; Programming Code: Java [.java] and Python [.py])
Download (1MB)

Abstract

The emerging Smart Grid (SG) paradigm promises to address decreasing grid stability from thinning safe operating margins, meet continually rising demand from pervasive high capacity devices such as electric vehicles (EVs), and fully embrace the shift towards green energy solutions. At the SG edge, widespread decentralisation of heterogeneous devices coupled with fluctuating energy availability and need as well as a greatly increased fluidity between their roles as energy producers, consumers, and stores raises significant challenges to ensuring robustness and security of both information and energy exchange. Detecting and mitigating both malicious and non-malicious threats in these environments is essential to the realisation of the full potential of the SG. To address this need for robust, localised, real-time security at the grid edge we propose CONCEDE, a collaborative cross-layer ego-network integrity awareness and attack impact reduction extension to our previous work on delay-tolerant cognitive adaptive energy exchange. We detail a substantial, targeted, energy disruption attack perpetrated by colluding mobile energy prosumers. Our CONCEDE proposal is then evaluated in multiple, diverse smart micro-grid (SMG) scenarios using hybrid traces of EVs and infrastructure from Europe, North America, and South America in the presence of a coordinated attack from malicious distributors seeking to disrupt energy supply to a target community. We show that CONCEDE successfully detects and identifies the nodes exhibiting malicious, dishonest behaviour and that CONCEDE also reduces the impact of a coordinated energy disruption attack on innocent parties in all explored scenarios across multiple criteria.

Item Type: Conference or Workshop Item (Paper)
Schools/Departments: University of Nottingham, UK > Faculty of Science > School of Computer Science
Depositing User: Radenkovic, Milena
Date Deposited: 15 Jan 2019 14:07
Last Modified: 07 Feb 2019 15:29
URI: https://eprints.nottingham.ac.uk/id/eprint/55772

Actions (Archive Staff Only)

Edit View Edit View