A Comparison of Optimal and Sub-optimal Iterative Equaliza- tion Techniques for Full-response CPM

It is well known that continuous phase modulation (CPM) serially concatenated with convolutional codes provides powerful error correction through iterative decoding. CPM also has the inherent benefits of being bandwidth efficient and compatible with non-linear amplifiers. A disadvantage of CPM is its relatively poor performance over ISI channels. In this paper we compare two methods of iteratively equalizing CPM. The first is the optimal method that combines the memory of the CPM with the memory of the channel such that decoding is performed over the resulting “super–trellis”. The second method is the parallel interference cancellation method using the Laurent signal model. The two methods are compared through Monte Carlo simulations and EXIT chart analysis.