A Survey and Comparison of Time-Delay Estimation Methods in Linear Systems

摘要: In this thesis the problem of time-delay estimation (TDE) in linear dynamic systems is treated. The TDE studied for signal-to-noise ratios, input signals, and that are common process industry. This also implies both open-loop closed-loop cases interest. true estimated, which may be different from giving best model approximation system. Time-delays not a multiple sampling interval interest to estimate.In thesis, review classification according underlying principles methods literature made. main classes are: 1) Time-Delay Approximation Methods: estimated relation (a model) between output signals expressed certain basis. an explicit parameter model. 2) Explicit Parameter 3) Area Moment integrals impulse step responses. 4) Higher Order Statistics Methods.Some new variants old ones suggested evaluated, some have good performance poor performance. Properties analyzed, theoretically experimentally. Recommendations given on how choose method signal. Generally, prediction error where optimized simultaneously with other parameters give quality.Most evaluations been conducted factorial experiments using Monte Carlo simulations open closed loop. Some statistical analysis utilized: RMS estimates gives absolute measure ANOVA (ANalysis Of VAriance) confidence intervals conclusions level confidence.