Symbol rate classification of PSK/QAM signals using direct frequency estimators

The flexible architecture provided by cognitive radio has spurned renewed interest in spectrum monitoring. To enable blind signal demodulation, the symbol rate must be determined. Recent techniques for determining symbol rates follow one of two broad approaches. The first approach estimates the carrier frequency and down-converts the signal for analysis at base-band. This approach is computationally expensive as carrier extraction hardware is needed for each modulation scheme. The other approach applies the wavelet transform to an over-sampled received signal to detect phase jumps and frequency changes. This paper introduces a new approach for detecting modulation-based carrier changes by monitoring the behaviour of an error signal associated with a direct frequency estimator. The proposed method has a simple architecture and eliminates the need for over-sampling seen in the wavelet transform approach. Symbol rate estimations were made under different carrier-to-noise ratios (CNRs) and modulation types and the results are provided.