Cost-Effective Proportionate Affine Projection Algorithm with Variable Parameters for Acoustic Feedback Cancellation

This paper focuses on mixed error cost function based Variable Step Size Improved Proportionate Affine Pro-jection Algorithm (VSS-IPAPA) with Prediction Error Method (PEM). The objective is to investigate the possible improvements in terms of performance, complexity and applicability in the real-life scenario of the updating algorithm. Mainly, the use of low complexity VSS Memory IPAPA (VSS-MIPAPA) and the addition of Variable Regularization Parameter (VRP) are proposed in the scope of this paper. As a result, VSS-MIPAPA-VRP is proposed as an improved version of the previous VSS-IPAPA algorithm. The simulation results demonstrate that VSS-MIPAPA-VRP achieves less time complexity, higher convergence accuracy and higher precision compared to the VSS-IPAPA.