Robust parameter tracking through regional forgetting

• Published in: 1995 International Conference on Acoustics, Speech, and Signal Processing Vol. 2; pp. 1440 - 1443 vol.2
• Main Authors: Shorten, R., Schutte, A., Fagan, A.D.
• Format: Conference Proceeding
• Language: English
• Published: IEEE 1995
• Subjects: Convergence; Digital signal processing; Educational institutions; Intelligent systems; Least squares methods; Linear regression; Output feedback; Power system modeling; Resonance light scattering; Robustness
• DOI: 10.1109/ICASSP.1995.480554
• ISBN: 0780324315; 9780780324312
• ISSN: 1520-6149

The recursive least squares (RLS) algorithm with exponential forgetting (/spl lambda/RLS) is perhaps the best known and most widely used algorithm for tracking the time varying parameters of a linear regression model. The implicit assumption in using the /spl lambda/RLS algorithm is that the information is uniformly distributed over the time horizon. Frequently this assumption does not hold and serious difficulties can be encountered when using many model structures. These include convergence of the parameters to local system or noise characteristics and output bursting, i.e. a large error when the operating point changes. In this paper several simple alternatives to the standard /spl lambda/RLS algorithm are proposed. The proposed algorithms extend the idea of a sliding window by quantising the whole input space. These algorithms considerably reduce the risk of forgetting useful information and eliminate the possibility of output bursting by relating the adaptation capabilities of the algorithm to the amount of input stimulation. Simulation results confirm the efficacy of our approach.