Estimation and Identification Process Using an Exponential Forgetting Factor

System identification and parameter estimation are important to obtain information from systems which are difficult to model and that are usually presented as BlackBox models. This work presents a point to point parameter estimation of a generalized non-deterministic system, whose results are variable through time, by using an exponential Forgetting Factor (FF). An average approximation is used as base to add an exponential FF to modify and improve the average results, without increasing the computational cost considerably. A comparison of the results applying the Least Square Method (LSM), the Recursive Least Square (RLS) and FF is presented using a signal for tracking a simple trajectory to prove the performance of the proposed method. As conclusion, it is obtained an online estimation for a non-deterministic signal without needing a previous training or Knowledge Base (KB).