International Journal of Industrial Electronics and Electrical Engineering(IJIEEE).
Paper Title - High Speed Recursive Noise Cancellation with Fast Convergence Digital System
The key objective of this paper is to provide an idea for VLSI Implementation of RLS algorithm for Noise
Cancellation with real time analog inputs. In this paper, we present an efficient architecture for the implementation of ANC
systems often for high-speed digital signal processors to cancel out disturbing noise. The throughput rate of the proposed
design is significantly increased by recursive update and concurrent implementation of filtering and weight-update
operations. The conventional LMS inner-product computation is replaced by conditional signed recursive accumulation in
order to reduce the sampling period and area complexity. The proposed implementation significantly outperforms the
existing implementations in terms of three important key metrics. 1. The least mean squares (LMS) algorithms adjust the
filter coefficients to minimize the cost function. Compared to least mean squares (LMS) algorithms, the RLS algorithms
achieve faster convergence by variable step size. 2. Proposed RLS algorithms require fewer computational resources and
memory than the RLS algorithms. 3. The implementation of the algorithms is less complicated due to lesser tap approach
than the all other existing algorithms. Through MATLAB simulation experiments efficiency of RLS over LMS will be
proved. The VLSI implementation results show that the proposed algorithm as superior performance in Fast convergence
rate, low complexity, and has superior performance in noise cancellation.
Keywords- Active noise cancellation(ANC), least mean square (LMS) ,recursive lease square(RLS)