Research Article
Predicting Radio Resource Availability in Cognitive Radio - an Experimental Examination
@INPROCEEDINGS{10.1109/CROWNCOM.2008.4562504, author={Shoji KANEKO and Shinichi NOMOTO and Tetsuro UEDA and Shingo NOMURA and Kazunori TAKEUCHI}, title={Predicting Radio Resource Availability in Cognitive Radio - an Experimental Examination}, proceedings={3rd International ICST Conference on Cognitive Radio Oriented Wireless Networks and Communications}, publisher={IEEE}, proceedings_a={CROWNCOM}, year={2008}, month={7}, keywords={cognitive radio optimal wireless system selection prediction radio resource availability AR model n-step-ahead prediction}, doi={10.1109/CROWNCOM.2008.4562504} }- Shoji KANEKO
Shinichi NOMOTO
Tetsuro UEDA
Shingo NOMURA
Kazunori TAKEUCHI
Year: 2008
Predicting Radio Resource Availability in Cognitive Radio - an Experimental Examination
CROWNCOM
IEEE
DOI: 10.1109/CROWNCOM.2008.4562504
Abstract
This paper presents a method for predicting radio resource availability in cognitive radio. In this paper, unlike a primary / secondary model, cognitive radio is defined as wireless communication technology in which each node communicates via an optimal wireless system based on recognition of radio resource availability in heterogeneous wireless communication systems. Therefore, it is important to be able to recognize such radio resource availability accurately. However, a cognitive radio node is unable to recognize the subsequent radio resource availability at the time of selecting the optimal wireless system. The authors focus on the prediction of radio resource availability to resolve the above issue. In this paper, the authors focus on IEEE802.11 [1] and radio channel occupation time, which is calculated from packet length and the transmission rate, is used as radio resource availability. From the results of the examination, it would seem feasible that the auto-correlation coefficient (or partial autocorrelation coefficient) could be utilized as information on the reliability of the prediction value when an auto-regression model (AR model) is used for the prediction. Furthermore, by comparing n-step-ahead prediction for a time series calculated for a 1-second interval and 1-step-ahead prediction for a time series calculated for an n-second interval, it is shown that the accuracy of the prediction in both cases is almost identical when the information volume to calculate a prediction expression is the same. Therefore, either prediction approach can be selected depending on the cognitive radio system.
