Generalized data-aided equalizers for digital communication channels

As the applications of digital communication techniques increase, demands for higher-data-rate communication and efficient spectrum utilization become greater. In particular, when digital techniques are used in high-speed, wireless mobile communications, intersymbol interference (ISI) becomes one of the fundamental performance limitations. Adaptive equalizers provide a general approach for which many algorithms and structures have been suggested over the last three decades to mitigate ISI limitation.; In this dissertation, new architectures of data-aided equalizers have been proposed. First, we propose the adaptive-canceller equalizer (ACE) structure. The basic principle of the ACE is that it decouples the feedforward and feedback filters. Because the response of the feedback filter is the estimated channel response, this decoupling allows the feedback filter to focus on the significant ISI components. Also, the selective subtraction of strong ISI components can achieve complexity reduction or better performance, particularly in finite impulse response implementations, than the decision-feedback equalizer (DFE) for channels which have long impulse responses.; We propose a method to achieve high-performance, pipelined, data-aided equalization using the ACE structure. Also, the ACE structure can be applied to the canceller structure. Converting the DFE structure to the canceller structure requires an extra equalizer, but the canceller structure of the ACE does not require extra filters. Since the feedback filter of the ACE is an estimation of the channel characteristic, only one feedback filter is required in the canceller structure of the ACE. It requires only an additional temporary decision device.; Although, the ACE structure has advantages in implementations of pipelined structure and canceller structure, we found that the performance of the ACE depends on the channel characteristics. Comparing the ACE with the DFE, the performance of the ACE is not always better than the performance of the DFE. Therefore, we propose a generalized data-aided equalizer that combines the structures of the ACE and the DFE in order to achieve performance and complexity optimization.