A Comparative Study of Anomaly Detection Techniques for Smart City Wireless Sensor Networks
AbstractIn many countries around the world, smart cities are becoming a reality. These cities contribute to improving citizens’ quality of life by providing services that are normally based on data extracted from wireless sensor networks (WSN) and other elements of the Internet of Things. Additionally, public administration uses these smart city data to increase its efficiency, to reduce costs and to provide additional services. However, the information received at smart city data centers is not always accurate, because WSNs are sometimes prone to error and are exposed to physical and computer attacks. In this article, we use real data from the smart city of Barcelona to simulate WSNs and implement typical attacks. Then, we compare frequently used anomaly detection techniques to disclose these attacks. We evaluate the algorithms under different requirements on the available network status information. As a result of this study, we conclude that one-class Support Vector Machines is the most appropriate technique. We achieve a true positive rate at least 56% higher than the rates achieved with the other compared techniques in a scenario with a maximum false positive rate of 5% and a 26% higher in a scenario with a false positive rate of 15%. View Full-Text
Share & Cite This Article
Garcia-Font, V.; Garrigues, C.; Rifà-Pous, H. A Comparative Study of Anomaly Detection Techniques for Smart City Wireless Sensor Networks. Sensors 2016, 16, 868.
Garcia-Font V, Garrigues C, Rifà-Pous H. A Comparative Study of Anomaly Detection Techniques for Smart City Wireless Sensor Networks. Sensors. 2016; 16(6):868.Chicago/Turabian Style
Garcia-Font, Victor; Garrigues, Carles; Rifà-Pous, Helena. 2016. "A Comparative Study of Anomaly Detection Techniques for Smart City Wireless Sensor Networks." Sensors 16, no. 6: 868.
Note that from the first issue of 2016, MDPI journals use article numbers instead of page numbers. See further details here.