Jump to main content
Jump to site search

Issue 10, 2014
Previous Article Next Article

Multi-criteria anomaly detection in urban noise sensor networks

Author affiliations

Abstract

The growing concern of citizens about the quality of their living environment and the emergence of low-cost microphones and data acquisition systems triggered the deployment of numerous noise monitoring networks spread over large geographical areas. Due to the local character of noise pollution in an urban environment, a dense measurement network is needed in order to accurately assess the spatial and temporal variations. The use of consumer grade microphones in this context appears to be very cost-efficient compared to the use of measurement microphones. However, the lower reliability of these sensing units requires a strong quality control of the measured data. To automatically validate sensor (microphone) data, prior to their use in further processing, a multi-criteria measurement quality assessment model for detecting anomalies such as microphone breakdowns, drifts and critical outliers was developed. Each of the criteria results in a quality score between 0 and 1. An ordered weighted average (OWA) operator combines these individual scores into a global quality score. The model is validated on datasets acquired from a real-world, extensive noise monitoring network consisting of more than 50 microphones. Over a period of more than a year, the proposed approach successfully detected several microphone faults and anomalies.

Graphical abstract: Multi-criteria anomaly detection in urban noise sensor networks

Back to tab navigation

Publication details

The article was received on 19 May 2014, accepted on 29 Jul 2014 and first published on 29 Jul 2014


Article type: Paper
DOI: 10.1039/C4EM00273C
Author version available: Download Author version (PDF)
Citation: Environ. Sci.: Processes Impacts, 2014,16, 2249-2258
  •   Request permissions

    Multi-criteria anomaly detection in urban noise sensor networks

    S. Dauwe, D. Oldoni, B. De Baets, T. Van Renterghem, D. Botteldooren and B. Dhoedt, Environ. Sci.: Processes Impacts, 2014, 16, 2249
    DOI: 10.1039/C4EM00273C

Search articles by author

Spotlight

Advertisements