An Eigenvalue Approach to Enhance Energy Detection in a Mobile Spectrum Monitoring Network

For fulfilling new tasks of the regulator, we propose a novel mobile spectrum monitoring network. The huge amount of measurement data contains besides thermal noise also noise induced by the Automatic Gain Control (AGC). As a result the occupancy results based on normal energy detection are significantly biased. In order to assess the spectrum occupancy more accurately, we employ an eigenvalue based technique to eliminate AGC noise components. This technique, using singular value decomposition, enables signal space analysis based on the received spectral data by finding an optimal threshold. In addition, this technique allows to eliminate non-linear noise induced by the AGC. In this paper we apply this method to the UMTS downlink band using collected data of a mobile measurement system. The results indicate that spectrum occupancy can be assessed 21% more accurate compared to an ITU-based method.