A low complexity detection algorithm for the primary synchronization signal in LTE -

Abstract

In 3GPP-LTE, upon call initiation a search procedure performed by the user equipment (UE) is triggered to synchronize to the transmitting base station also referred to as eNodeB and obtain the frame synchronization, the symbol timing, the carrier frequency offset, and the cell identity information. Frame synchronization is used to determine the beginning of each frame and has to work continuously since the UE might be in continuous mobility, so it has to periodically look up for neighboring cells for possible handovers. Our concern is to design a fast frame synchronization algorithm that is computationally efficient. Efficient computation is achieved by reducing the number of multiplications which lead to reduce the power consumption in mobile batteries along the continuous frame synchronization procedure. However the proposed algorithm in this thesis doesn't trade the performance for complexity reduction. Good performance corresponds to acquiring the signal in a short period of time and maintaining immunity to the carrier frequency offset. Present algorithms proposed in the literature are based on correlation methods that involve a large number of multiplications and thus lead to high power consumption. In this thesis a novel algorithm is proposed that reduces the complexity of the primary synchronization signal (PSS) frame synchronization in the downlink from 24 complex multiplications based on the current known literature to just 6. Instead of working on the properties of the PSS and the LTE frame structure to reduce the complexity we generalize the problem to be an optimization problem, so we propose a clustering algorithm based on Lloyd algorithm to partition the synchronization signal into 8 sets where each set is represented by a leader, and hence the corresponding PSS take only 8 distinct values, and the number of complex multiplication is reduced from 24 to 8. In addition a special manipulation of the clustered PSS allows us to reduce the complexity furthermore to 6 complex multiplications. The proposed