A novel adaptive iterative detection technique for joint estimation and detection

Reliability of the communication systems depends hugely on the receiver performance where the synchronization and detection tasks need to be performed. Classically these two tasks are attended separately resulting in simple yet non-optimal receivers. During the last decade, a family of iterative receivers has been introduced to approximate the optimal solution to joint estimation and detection problem by means of the turbo principle. With the help of graph theory and the belief propagation framework these receiver structures are unified and soft information driven schemes emerged, leading to more reliable detection. The essence of these schemes lays in factorization of a global function whose marginal corresponds to the objective function and obtained via simple message passing algorithms. On the other hand, as far as our knowledge, all of the proposed structures pursue a specific factorization while devising their respective schemes. In this work utilizing a different factorization, we introduce a novel iterative receiver for joint equalization/detection problem. With the aid of Variational Bayesian approximation we show that the complexity can be reduced without sacrificing the error performance drastically