EVM Based Primary User Monitoring in Cognitive Radio Systems

Performance of quickest spectrum sensing for EVM‐based change detection

“…Hence, it is of crucial importance to design a system that can detect the incumbent PU efficiently and robustly. This process of detecting the PU during SU transmission is termed spectrum monitoring which has been noted in [3–6]. The method in [3] uses periodic spectrum sensing where SU transmission is switched off every once in a while to create a silent sensing interval for probing the channel which makes it spectrally inefficient and results in a sensing throughput tradeoff.…”

“…The method in [3] uses periodic spectrum sensing where SU transmission is switched off every once in a while to create a silent sensing interval for probing the channel which makes it spectrally inefficient and results in a sensing throughput tradeoff. The spectrum monitoring methods employed in [4–6] are done on the SU receiver during the reception of packets, so that communication can occur at the same time as monitoring is performed and no quiet period is required as in [3]. The idea in [4] is to compare the bit error count for each received packet to a threshold value.…”

“…However, as certain subcarriers are always reserved for spectrum monitoring, it degrades SU throughput. In [6], we have proposed a spectrum monitoring method for an OFDM based CR system based on the error vector magnitude (EVM) metric that can be obtained immediately during demodulation and decoding of the received packets. Our system does not impact on the SU throughput performance as we use only the inherent pilots of the OFDM systems.…”

“…All of the spectrum monitoring methods mentioned in [3–6] focus on maximising probability of detection subject to a fixed false alarm rate. When the PU reappears, the SU needs to detect its transmission quickly to vacate the channel and when the PU stops transmission, the SU must be able to detect the absence of the PU quickly.…”

“…We utilise the EVM of the received signal, which is typically calculated in any receiver as a part of normal baseband signal processing for channel estimation and soft decision decoding. Earlier in our work on [6], we have analysed EVM‐based detection with a fixed sample size. There, a block of samples is taken, the EVM test statistic is calculated and subsequently a decision is made by comparing the value of the test statistic to a predetermined threshold.…”