Minimax long range parameter estimation

Johnson Tse, Joseph Bentsman, Norman Miller

Research output: Contribution to journalConference articlepeer-review


This paper derives a robust long range prediction error estimator. The long range minimax prediction error (MPE) algorithm is developed by combining the robust (H) linear predictor and long range prediction error method. The resulting identification algorithm, which minimizes the peaks of the error spectrum rather than its integral on the unit circle, is shown to have better robustness properties than recursive least squares with ad hoc data prefiltering. The frequency domain properties of the MPE estimator indicate that the MPE estimator is equivalent to the RLS algorithm with prefilter. The MPE estimator can provide a better estimated model for the model based long range predictive controller. The resulting self-tuning predictive controller will have better overall stability and robustness properties then the H2 self-tuning algorithms.

Original languageEnglish (US)
Pages (from-to)277-282
Number of pages6
JournalProceedings of the IEEE Conference on Decision and Control
StatePublished - 1994
EventProceedings of the 33rd IEEE Conference on Decision and Control. Part 1 (of 4) - Lake Buena Vista, FL, USA
Duration: Dec 14 1994Dec 16 1994

ASJC Scopus subject areas

  • Control and Systems Engineering
  • Modeling and Simulation
  • Control and Optimization


Dive into the research topics of 'Minimax long range parameter estimation'. Together they form a unique fingerprint.

Cite this