Optimized LMS algorithm for system identification and noise cancellation

Optimization by definition is the action of making most effective or the best use of a resource or situation and that is required almost in every field of engineering. In this work, the optimization of Least Mean square (LMS) algorithm is carried out with the help of Particle Swarm Optimization (PSO) and Ant Colony Optimization (ACO). Efforts have been made to find out the advantages and disadvantages of combining gradient based (LMS) algorithm with Swarm Intelligence SI (ACO, PSO). This optimization of LMS algorithm will help us in further extending the uses of adaptive filtering to the system having multi-model error surface that is still a gray area of adaptive filtering. Because the available version of LMS algorithm that plays an important role in adaptive filtering is a gradient based algorithm, that get stuck at the local minima of system with multi-model error surface considering it global minima, resulting in an non-optimized convergence. By virtue of the proposed method we have got a profound solution for the problem associated with system with multimodal error surface. The results depict significant improvements in the performance and displayed fast convergence rate, rather stucking at local minima. Both the SI techniques displayed their own advantage and can be separately combined with LMS algorithm for adaptive filtering. This optimization of LMS algorithm will further help to resolve serious interference and noise issues and holds a very important application in the field of biomedical science.