A Robust Family of Algorithms for Adaptive Filtering Based on the Arctangent Framework

This brief introduces a novel cost function framework for developing robust algorithms for adaptive filtering by embedding the standard cost function into the arctangent framework. This proposed framework is called the arctangent cost function framework. Based on this, we propose an arctangent family of robust algorithms for adaptive filtering. The performance of the proposed family of algorithms is tested through simulation studies in system identification scenarios that confirm the enhanced performance achieved by the arctangent family of algorithms over standard algorithms.