Development of Efficient Algorithms for Impulsive Noise Reduction

Abstract

Impulsive noise is a category of acoustic noise which includes unwanted, instantaneous sharp sounds. Impulsive noise is a manmade noise and has catastrophic effects in high data rate applications and communication systems. All the non-Gaussian noises come under the category of impulsive noise. Impulsive noise produces burst errors, as a result information is significantly corrupted and band width is wasted. In this thesis we have proposed efficient algorithms for reduction of impulsive noise in OFDM and ANC realm. All the proposed algorithms are based on adaptive filtering philosophy. In the first part of this research, State Space Recursive Least Square (SSRLS) algorithm based enhanced impulsive noise canceler is proposed. The suggested algorithm was tested on sinusoidal and Electrocardiogram signal and was successful in significantly reducing the impulsive noise. The same impulsive noise canceller again outperformed the existing techniques when it was implemented in OFDM system. Second part of the thesis presents a new hybrid dual facetted technique for impulsive noise suppression in OFDM systems employing error correction code (Reed Solomon) and adaptive filters. Adaptive filtering achieves more accurate estimate of the original OFDM signal after impulsive noise cancellation. The results in terms of steady state mean square error (MSE) reduction, bit error rate (BER) improvement and signal to noise ratio (SNR) enhancement confirm the effectiveness of the proposed hybrid approach when compared with the recently reported techniques. II Moreover, a Filtered-x SSRLS (FxSSRLS), an SSRLS based practical adaptive solution for Active Noise Control (ANC) is suggested in this dissertation. The proposed FxSSRLS algorithm is more robust in eliminating high-peaked impulses than the recently reported algorithms for ANC applications. Moreover, the suggested solution exhibits better stability and faster convergence, without jeopardizing the performance of the proposed solution in terms of residual noise suppression in the presence of impulses. Last but not the least, another efficient Filtered x Bhagyashri (FxBhagyashri) algorithm for impulsive noise mitigation in ANC is also proposed. FxBhagyashri algorithm becomes unstable in presence of impulsive noise, so two modifications in the FxBhagyashri algorithm i.e; Clipped FxBhagyashri (CFxBhagyashri) and Modified Filtered x Bhagyashri (MFxBhagyashri) algorithms are also presented in this research manuscript. Both modifications gave better results in impulsive noise reduction as compared to standard FxBhagyashri algorithm. It was also found that the proposed MFxBhagyashri algorithm can approach FxRLS algorithm in term of low steady state error with almost same computational complexity of FxLMS family algorithms. In the end of dissertation, the closed form expression for steady state analysis of Normalized Bhagyashri algorithm is presented on same lines as of Bhagyashri algorithm which serve as a motivation for the researchers who intend to work in this area.