An Adaptive Energy Detection Scheme with Real-Time Noise Variance Estimation

Energy detection-based spectrum sensing techniques are ideally suited for power-constrained cognitive radio applications because of their lower computational complexity compared to feature detection techniques. However, their detection performance is dependent on multiple factors like accuracy of noise variance estimation and signal-to-noise ratio (SNR). Many variations of energy detection techniques have been proposed to address these challenges; however, they achieve the desired detection accuracy at the cost of increased computational complexity. This restricts the use of enhanced energy detection schemes in power-constrained applications such as aeronautical communication. In this paper, an adaptive low-complexity energy detection scheme is proposed for spectrum sensing in an L-band Digital Aeronautical Communication System (LDACS) at lower SNR levels. Our scheme uses a real-time noise variance estimation technique using autocorrelation which is induced by the cyclic prefix property in LDACS signals. The proposed technique does not incur dedicated hardware blocks for noise variance estimation, leading to an efficient hardware implementation of the scheme without significant resource overheads. The simulation studies of the proposed scheme show that the desired accuracy (90% detection accuracy with only 10% of false alarms) can be achieved even at $$-16.5$$ - 16.5  dB SNR, significantly lowering the SNR wall over existing energy detection schemes.