A design of energy detector in cognitive radio under noise uncertainty

Cognitive radio is proposed as an approach to solve the conflict between the spectrum congestion and under-utilization. Energy detector is the detector most widely used in cognitive radio system due to its generality and low complexity. In this paper, the performance of energy detector under noise uncertainty is studied. The discrete form and continuous form of the noise uncertainty model are proposed. The average detection performances of the energy detector are analyzed under various types of noise uncertainty models, based on which a design of energy detector is put forward. Selecting different statistic decision thresholds will result in different detection performance. According to a prior knowledge about the noise variance distribution, a proper threshold can be selected to improve the detection performance as much as possible under the existence of noise uncertainty.