Cognitive Radio

The radio frequency spectrum is a finite natural resource, and its growing demand from wireless communication systems has made efficient utilization a critical engineering challenge. Cognitive radio is an intelligent wireless communication system that can sense its surrounding radio environment and dynamically adapt its transmission parameters to use available spectrum without interfering with licensed users.

Core Concept Explanation

Traditional spectrum allocation assigns fixed frequency bands to specific services through regulatory bodies. Studies have consistently shown that large portions of licensed spectrum remain unused at any given time and location. This underutilization is called spectral inefficiency. Cognitive radio addresses this by allowing unlicensed secondary users to opportunistically access these unused bands called spectrum holes or white spaces.

The primary user (PU) is the licensed owner of a frequency band and has priority of access at all times. The secondary user (SU) is the cognitive radio device that may use the spectrum only when the primary user is absent. The moment the primary user returns and starts transmitting, the secondary user must immediately vacate that band without causing interference.

The four key functions of a cognitive radio system are: spectrum sensing (detecting who is using which frequency), spectrum decision (selecting the best available hole), dynamic spectrum access (DSA) (actually transmitting in the selected hole), and spectrum mobility (switching to another hole if the primary user reclaims the current one). These together form the cognitive cycle.

Spectrum Sensing Techniques

The most fundamental task in cognitive radio is detecting whether a primary user is present on a given channel. Energy detection is the simplest method: the cognitive radio measures the received signal energy over a frequency band and compares it against a threshold. If energy exceeds the threshold, the band is marked as occupied. It works without knowing the primary user signal structure but is sensitive to noise uncertainty.

Matched filter detection is more accurate and requires prior knowledge of the primary user signal format. It cross-correlates the received signal with a known template to detect presence. Cyclostationary feature detection exploits the periodicity embedded in communication signals such as pilot tones or cyclic prefixes. It is more robust in low SNR environments but computationally expensive.

Mathematical Expression

Energy detection uses a binary hypothesis test. Under hypothesis H0 (PU absent), the received signal is only noise: r(t) = n(t). Under hypothesis H1 (PU present), the received signal contains both signal and noise: r(t) = s(t) + n(t). The test statistic for energy detection over N samples is expressed as the sum of squared samples divided by the noise variance. The detection probability Pd and the false alarm probability Pfa are the two key performance parameters. A high Pd means PU is reliably detected. A low Pfa means the SU rarely mistakes noise for a PU signal, preserving spectrum access opportunities.

The sensing threshold lambda determines the trade-off between Pd and Pfa. Increasing lambda decreases Pfa but also decreases Pd, creating a receiver operating characteristic (ROC) curve. For GATE purposes, remember that the complementary relationship between these two probabilities and the effect of SNR on detection performance are the most tested concepts.

Practical Understanding

Cognitive radio is the enabling technology for dynamic spectrum management in 5G and beyond networks. The TV white space (TVWS) bands in the UHF spectrum are a practical example where cognitive radio devices operate in unused TV channels. The IEEE 802.22 standard defines a wireless regional area network that uses cognitive radio to operate in TVWS bands.

The two primary performance concerns in practical deployment are the hidden primary user problem, where the SU cannot hear the PU due to deep fading or shadowing but the PU's receiver is still being interfered with, and the reporting channel reliability in cooperative sensing. Cooperative spectrum sensing involves multiple secondary users sharing their sensing information to arrive at a combined decision, improving overall detection accuracy.

Given:
Noise variance = 1, SNR of PU signal = 5 dB = 3.162 (linear), N = 100 samples, threshold lambda = 120

Why this formula applies:
Energy detection test statistic T = sum of r^2[n] for n=1 to N. Under H0, T ~ Chi-squared with 2N degrees of freedom scaled by noise variance.

Formula:
Mean of T under H0: mu_0 = N * sigma_n^2 = 100 * 1 = 100
Mean of T under H1: mu_1 = N * (sigma_s^2 + sigma_n^2) = 100 * (3.162 + 1) = 416.2

Substitution:
Threshold lambda = 120
SNR = 5 dB means signal power = 3.162 * noise power

Calculation:
Normalized threshold under H0: (120 - 100) / sqrt(2*100) = 20 / 14.14 = 1.414
Pfa = Q(1.414) = Q(sqrt(2)) ≈ 0.079 or approximately 7.9%

Final Answer:
False alarm probability Pfa ≈ 7.9%, meaning about 8% of the time the SU incorrectly senses the band as occupied when PU is absent.

Exam Tip: GATE often tests the trade-off between Pd and Pfa. Remember increasing the threshold lambda reduces false alarms but also reduces the probability of detecting the primary user, which is the dangerous case. Also remember cognitive radio secondary users must always vacate when PU returns, not just when sensing the PU.

• Energy detection measures received power and compares with a threshold. It is the simplest but most noise-vulnerable sensing method.
• Matched filter detection maximizes SNR by correlating received signal with known PU waveform, requiring signal structure knowledge.
• Cyclostationary detection uses periodic statistical features of communication signals, effective at low SNR.
• Cooperative sensing combines decisions from multiple SUs to resolve the hidden PU problem caused by shadowing or fading.
• DSA allows SU to dynamically select, use, and switch frequency bands based on real-time spectrum availability.

Quick Revision

• Cognitive radio enables secondary users to access licensed spectrum during primary user inactivity without causing interference.
• Spectrum hole (white space): a frequency band that is licensed but currently unused at a specific time and location.
• Cognitive cycle: Sense spectrum, analyze and decide, access the hole, vacate when PU returns.
• Energy detection test statistic: T = (1/N) * sum[r^2(n)], compared against threshold lambda.
• Pd = probability of correctly detecting PU presence. Pfa = probability of falsely declaring PU present when absent.
• Increasing threshold lambda: Pfa decreases, Pd decreases. Decreasing lambda: Pfa increases, Pd increases.
• Hidden PU problem occurs when SU cannot detect PU due to multipath or shadowing. Cooperative sensing is the solution.