Iterative envelope-based SNR estimation for nonconstant modulus constellations

The signal-to-noise ratio (SNR) estimation problem is addressed for linearly modulated signals over flat-fading channels. A new SNR estimator based on the expectation-maximization algorithm is proposed in order to iteratively maximize a certain approximation to the envelope-based likelihood function. The new method is robust to carrier offsets and can be applied to nonconstant modulus constellations. It operates in a non-data-aided manner, although it can be suitably modified to exploit the availability of pilot symbols within the frame.