A Maximum Likelihood Approach for SSS Detection in LTE Systems

Before establishing a communication link with the serving base station (eNodeB), user equipment (UE) operating in a long-term evolution (LTE) multi-cellular network must acquire some specific information, including the sector identity and cell group identity. For this purpose, two training sequences called primary synchronization signal (PSS) and secondary synchronization signal (SSS) are periodically transmitted in the downlink to convey such information. In this work, we present a novel maximum likelihood (ML) approach for SSS detection assuming that the PSS has been successfully identified at an earlier stage. As we shall see, the resulting scheme turns out to be too complex for practical implementation as it requires perfect knowledge of the channel covariance matrix. Therefore, we look for simpler solutions and propose two reduced-search methods that operate in a mismatched mode. The first scheme exploits channel state information emerging from both the primary and secondary synchronization signals, while the second scheme operates using only the secondary synchronization signal. Numerical analysis indicates that the proposed methods outperform existing alternatives and can be successfully applied even in a severe propagation scenario. The price for such an advantage is a certain increase of the processing requirement.