Communications have achieved great development on several fronts over the years. Among these, the communication using chaotic signals has been object of growing interest among researchers due to the characteristics of spread spectrum and hard detection. However these techniques still have inferior performance in comparison to conventional methods in non-ideal channels. To contribute do the solution of this problem, statistical estimation algorithms have been applied to the detection of the transmitted signal. The objective of this thesis is to study a communication system using chaotic carriers and reception with maximum likelihood (ML-CSK - Maximum Likelihood Chaos Shift Keying). For this, the application of Viterbi algorithm in chaotic modulation signals is investigated. As the previously proposed algorithms offer only good performance to signals generated by maps that present well-behaved probability density, a new technique was designed based on analysis of the map characteristics, obtained numerically through a training vector. The results of performed simulations assure the applicability and the good performance of the proposed innovations.
