Selection of multi-h modulation sets using a stochastic search algorithm

Multi-h CPFSK modulation is simple to implement, spectrally efficient, and has inherent error correction properties. Multi-h requires large parameter sets to obtain high coding gain. Large numbers of combinations may need to be investigated. A fast efficient and thorough search needs to be performed. A search method based on a genetic algorithm known as `population based incremental learning´ (PBIL) is described. A number of sets of modulation indices are shown and their performance is evaluated