Research On Narrowband Interference Rejection For Power-Line Communication Systems

Recently, broadband Power-line Communication (PLC) has been a hot research topic in the world with its advantages of wide coverage, no wire line and convenient connection and is expected to be the solution of the "last mile" for the Information Superhighway. However, power-line channel is plagued with numerous difficulties, including multipath distortion, narrowband interference (NBI) and impulsive noise. Thus, how to avoid NBI has been the first and foremost problem to be settled in the popularization of PLC access.Based on Interference Suppressing-Orthogonal Frequency Division Multiplexing (IS-OFDM) technology, this paper focus on NBI suppressing for the broadband PLC systems. There are two innovations in the dissertation. (1) A universal decoding method for both real and complex systems is presented (2) A new algorithm called cyclic prefix-recursive least square (CP-RLS) algorithm is also proposed.Compared to the originally OFDM, IS-OFDM has the capacity of suppressing NBI by spreading the signal power to other sub-carriers using orthogonal Hadamard. In IS-OFDM system, the output signal of Inverse Fast Fourier Transform (IFFT) is real owing to its coding conjugate symmetrically. Thus, it is a real transmission system. In the decoding, the first half of the fast Fourier transform (FFT) output signal is taken directly as the decoding result owing to the transmitting signal is real and the output signal of FFT is conjugate symmetric. Consequently, the decoding method of IS-OFDM is suitable for real system other than complex system. For this reason, a universal decoding method suitable for both real and complex system is proposed and the effectiveness of which is proved by simulation.However, the capacity of suppressing NBI for IS-OFDM is too limited to meet the requirements. For further improvement, adaptive filter technology is used in conjunction with IS-OFDM and two algorithms called Cyclic Prefix-Block Least Mean Square (CP-BLMS) and Linear Fast Block Least Mean Square are proposed.As is well known, Least Mean Square (LMS) algorithm and Recursive Least Square (RLS) algorithm are the most classical methods in the adaptive filter theories. Both CP-BLMS and LFBLMS algorithms are belong to the category of LMS. By comparison, in this paper a new algorithm called Cyclic Prefix-Recursive Least Square (CP-RLS) which belongs to the category of RLS is proposed. Simulation results indicate that the bit-error-rate (BER) performance of CP-RLS algorithm is not as good as that of CP-BLMS and FBLMS in low-power NBI scenario. On the contrary, the BER performance of the proposed CP-RLS algorithm obtains an order of magnitude improvement in the high-power NBI scenario and remains its steady performance with the increase of power of the NBI. The proposed CP-RLS algorithm is more suitable for PLC systems whose channel environment is noisy and troublesome.