Tek antenli haberleşme alıcıları için girişim giderimi yaklaşımı

In communication receivers, the system capacity and quality is often degraded by the interference signal. Even though antenna arrays have been the common method to separate interference signal, they may not be feasible for the receivers especially in the direction of down-link because of the physical and economic limitations. Therefore smart algorithms for interference cancellation have been developed that are based only on one received signal delivered by a single receive antenna. In this thesis, we proposed a complete solution that addresses the single-antenna receiver interference scenario with two co-channel users using π/2-BPSK modulation scheme. We divided the proposed solution into cascade of different algorithm blocks for performing distinct tasks. First block targets to estimate the channel corresponding to the initial user, the one who established the first communication link, within a predefined time interval in which most probably interference signal has not been received. The next block monitors the receive signal to detect the start of the interference event. We proposed alternative approaches to detect interference event based on the signal statistics. Once the interference event is detected the third block which is responsible for separating interference from the first user is activated. As a part of the algorithm for this block, we proposed a transformation that converted complex receiver samples to two real signals so that 2×1 complex under-determined source separation problem is converted to 2×2 real determined source separation problem. The main algorithm for the third block is based on Bounded Component Analysis (BCA) proposed as an efficient solution to 2 × 2 blind source separation problem obtained by the afore-mentioned transformation. The basic merit of this approach is its low complexity and its ability to achieve high separation performance for relatively low number of samples. The final block is responsible for detecting relative phase offset between sources as well as their symbols based on the Maximum Likelihood approach. The proposed blocks were implemented and tested in both floating point and fixed point MATLAB implementation. We also proposed VHDL implementation guidelines for the hardware realization of the overall system.