REM based approach for hidden node detection and avoidance in cognitive radio networks

This paper examines the benefit of using a Radio Environment Map (REM) exploiting measurement based signal level prediction as a means to detect and avoid the hidden node problem within wireless networks. Practical implementation considerations are taken into account, such as support for multi-Radio Access Technology (RAT) deployments, unknown transmit power levels, radio measurement inaccuracies and propagation anomalies (such as fading), which can reduce the reliability of REM based predictions. The approach considered within this paper manages to overcome these limitations and the expected performance is assessed in a representative deployment scenario. The novelty of the proposed approach lies in the ability to combine measurements to perform signal localization and a predicted signal space model using measurements taken by a grid of sensors. The model is then used for prediction of hidden node conditions and can be exploited in existing and future networks employing carrier sensing based medium access control or dynamic frequency re-use, or to other radio interference and cognitive radio related problems.