Wireless Spectrum Occupancy Prediction Based on Partial Periodic Pattern Mining

Cognitive radio appears as a promising technology to allocate wireless spectrum between licensed and unlicensed users in an efficient way. The availability of spectrum holes vastly affects the throughput and delay of unlicensed users. Predictive methods for inferring the availability of spectrum holes can help to improve channel utilization and reduce collision rate. In this paper, a spectrum occupancy prediction method based on Partial Periodic Pattern Mining (PPPM) is introduced. The mining aims to identify frequent occupancy patterns that are hidden in the spectrum usage of a channel, and then the mined frequent patterns are used to predict future channel states. By further extending our three states PPPM to N-states PPPM, the duration of high/low utilization on a channel is also predicted. The frequent patterns of channel utilization duration are critical in optimizing channel switching strategies. PPPM outperforms traditional Frequent Pattern Mining (FPM) by considering patterns that may not repeat perfectly due to noise, sensing errors, and irregular behaviors. Using real life network activities we show a significant reduction in miss rate. In addition, we observed that distinguishing low utilization periods from high utilization periods and mining rules in corresponding utilization periods significantly improve the prediction performance. With prediction mechanism, we show the performance of dynamic spectrum access is substantially improved. The high accuracy of duration prediction is also validated with data collected in the paging bands.

 by Pei Huang, Chin-Jung Liu, Li Xiao, Jin Chen, proceedings of IEEE / ACM 20th Intl Workshop on Quality of Service (IWQoS), Coimbra, Portugal, Jun. 2012