One of most the important problem in digital communication systems is the inter symbol-interference effect which arises from the frequency selective property of the communication channel. Inter symbol-interference effect is the most important problem that constrain high speed data transmission. Generally there are two solution methods to solve this problem. One of which is equalizer filter and the other one is maximum likelihood sequence estimation. Since computational complexity of the maximum likelihood sequence estimation is high, equalizer filters have become an important component for communication systems. In most of the current communication systems, the training sequence is utilized for channel estimation and equalization. This situation reduces the effective data rates and bandwidths of the communication systems. At the same time, usage of the training sequences should be avoided for high speed data communications. It is clear that high speed data rates and bandwidth efficiency need to be supported by the next communication standards. For this reason, a communication system without training sequence is needed. In this thesis, a solution based on the channel matched filter for blind channel equalization problem is presented. In order to improve the performance of the system, especially at the convergence speed, the swarm intelligence is applied to the proposed method. According to obtained MSE and BER results, the proposed method can be an alternative to the non-blind systems in real-time applications.
Key Words: blind channel estimation, blind channel equalization, channel matched filter, swarm intelligence, particle swarm optimization, turbo codes.