A Compact and High-Performance Acoustic Echo Canceller Neural Processor Using Grey Wolf Optimizer along with Least Mean Square Algorithms

Recently, the use of acoustic echo canceller (AEC) systems in portable devices has significantly increased. Therefore, need for superior audio quality resource-constrained opens new horizons creation high-convergence speed adaptive algorithms and optimal digital designs. Nowadays, AEC mainly least mean square (LMS) algorithm, since its implementation hardware architectures demands low area consumption. However, performance cancellation is limited. In addition, this algorithm presents local convergence optimization problems. approaches, based on stochastic algorithms, have emerged to increase probability encountering global minimum. simulation these requires high-performance computational systems. As a consequence, only been conceived as theoretical approaches. low-complexity potentially allows development compact architectures. paper, we propose convex combination, grey wolf LMS save achieve high by exploiting maximum best features each algorithm. proposed combination shows tracking capabilities when compared with existing Furthermore, present neuromorphic architecture simulate combination. Specifically, customized time-multiplexing control scheme dynamically vary number search agents. To demonstrate architecture, performed exhaustive testing. way, proved that it can be used real-world scenarios.