Centralized Integrated Spectrum Sensing for Cognitive Radios

Spectrum sensing is the challenge for cognitive radio design and implementation, which allows the secondary user to access the primary bands without interference with primary users. Cognitive radios should decide on the best spectrum band to meet the Quality of service requirements over all available spectrum bands. This paper investigates the integrated centralized spectrum sensing techniques in multipath fading environment and the performance was analyzed with energy detection and wavelet based sensing techniques for unknown signal. Keywords: Cognitive Radio, Spectrum Sensing, Signal Detection, Primary User, Secondary User

💡 Research Summary

The paper addresses the fundamental challenge of spectrum sensing for cognitive radio (CR) systems, focusing on a centralized, integrated approach that operates effectively in multipath fading environments. The authors begin by highlighting the under‑utilization of allocated spectrum—up to 70 % of the band may be idle at any given time—and cite the FCC’s recommendation to improve spectral efficiency through dynamic spectrum access. In this context, CR devices must reliably detect the presence of primary users (PUs) to avoid harmful interference while opportunistically exploiting vacant channels.

A centralized cooperative sensing architecture is proposed. Multiple secondary nodes collect raw received samples and forward them to a fusion center, which performs a unified detection process. This design overcomes the practical difficulty of direct channel measurement between a primary transmitter and a secondary receiver, and it leverages spatial diversity to improve detection reliability.

The core contribution is a bandwidth‑driven selection of two complementary detection techniques. For unknown narrow‑band signals, the classic energy detector (ED) is employed. The received signal is first filtered by a band‑pass filter (BPF) to isolate the band of interest, then squared and integrated over an observation window. The resulting energy is compared against a threshold determined by a constant false alarm rate (CFAR) criterion. Spectral estimation is carried out using FFT‑based periodograms, with modified periodogram methods employed to reduce variance and bias. Simulation results show that, in fading channels, the ED achieves a detection probability (Pd) greater than 0.9 when the signal‑to‑noise ratio (SNR) exceeds –5 dB, while maintaining a false alarm probability (Pfa) around 10⁻³. The trade‑off between missed detections and false alarms is illustrated, confirming that the optimal threshold occurs when Pd ≈ Pfa.

For wide‑band signals, a wavelet‑based detector is introduced. The method models the entire spectrum as a train of consecutive sub‑bands; within each sub‑band the power spectral density (PSD) is smooth, but abrupt changes occur at sub‑band boundaries. By applying a discrete wavelet transform to the PSD of the observed signal, singularities corresponding to vacant frequency bands are identified. A multiscale product technique is used to suppress spurious extrema caused by noise, improving robustness. The authors acknowledge that high sampling rates are required to capture the full bandwidth, and that inter‑system interference can degrade performance when multiple CR networks coexist. Nevertheless, the wavelet detector successfully reveals vacant bands in the simulated wide‑band scenario.

The integrated framework operates as follows: the fusion center first estimates the spectral width of the incoming signal. If the width falls below a predefined narrow‑band threshold, the ED is invoked; otherwise, the wavelet detector is applied. This adaptive strategy maximizes spectral efficiency while keeping computational complexity manageable. Experimental results demonstrate that the hybrid approach yields higher overall spectrum utilization compared with using either detector alone.

In conclusion, the study validates that a centralized, bandwidth‑adaptive sensing scheme can effectively protect primary users and enhance secondary access in realistic fading channels. The key insights include: (1) the importance of cooperative data collection for reliable detection, (2) the necessity of CFAR‑based threshold setting for balancing Pd and Pfa, (3) the advantage of wavelet multiscale analysis for wide‑band vacancy identification, and (4) the practical trade‑offs involving sampling rate and hardware complexity. Future work is suggested on low‑cost high‑speed ADC design, interference mitigation among multiple CR networks, and real‑time implementation of the proposed algorithms.