Cochannel interference (CCI) is a major impairment in narrowband cellular systems. To increase the spectral efficiency of the narrowband systems, identical carrier frequencies are reused in distant cells. The interference rejection capability of the receiver determines this frequency reuse and is therefore critical. In this thesis, we propose an improved demodulation scheme, employing high-resolution frequency estimation techniques, for continuous phase modulated (CPM) signals in presence of CCI. Minimum shift keying (MSK), which is a special case of CPM, is a very popular modulation format around the world. Frequency detectors, such as the limiter-discriminator permit the non-coherent demodulation of MSK signals. High-resolution frequency estimation appears as a very attractive alternative to the conventional non-coherent frequency detectors. The frequency estimation methods that we have studied are based on autoregressive modeling. The contributions of this thesis include the implementation of various demodulation schemes employing parametric frequency estimation. The use of the Viterbi algorithm as a non-linear equalization technique to mitigate intersymbol interference is considered. We verified that the model-based sequence estimation schemes outperform the conventional non-coherent receivers for MSK with AWGN, flat fading, and CCI. Demodulator diversity is also investigated as a way to combat interference. An improved technique combining the proposed model-based receiver and the conventional coherent receiver is implemented and simulated in presence of CCI. / Master of Science