An iterative detection algorithm for coded CPFSK signals with irrational modulation index

An algorithm for iterative demodulation and decoding of convolutionally coded and interleaved Continuous Phase Frequency Shift Keying (CPFSK) signals with an irrational modulation index is presented. The algorithm is based on integrating the Per-Survivor Processing (PSP) technique into a Soft-Input Soft-Output (SISO) CPFSK demodulator. Using PSP, the phase offsets emerging from the difference between the irrational transmit modulation index and the employed rational receive modulation index are estimated within the PSP-SISO algorithm at each state and in each time step. These phase offsets are taken into account to calculate the soft information. The phase offsets and the soft information are then refined during several iterations and used to estimate the transmit symbols. Simulations have shown that with only five trellis phase states and after only four iterations, the proposed algorithm performs at BER=10^-5 only 0.5 dB worse than a quasi-exact algorithm which has sixteen trellis phase states. This enables a significant reduction in computational complexity in terms of the number of the trellis phase states.