Deconvolution of Fast Pulses

Deconvolution of Fast Electrical Pulses for Standard Purposes
Típus: 
Nemzetközi (egyéb)
Kezdés éve: 
1996
Befejezés éve: 
1997
Partnerek: 
Megbízó: National Institute of Standards and Technology (NIST), USA

Tanszéki projektvezető

A munkatárs fényképe
head of department, habilitated associate professor
Szoba: IE442
Tel.:
+36 1 463-2065
Email: daboczi (*) mit * bme * hu

Tanszéki résztvevők

A munkatárs fényképe
head of department, habilitated associate professor
Szoba: IE442
Tel.:
+36 1 463-2065
Email: daboczi (*) mit * bme * hu

Contact information

Koordinátor: 
Dabóczi Tamás, BME MIT
Felelős: 
Tamás Dabóczi

Bemutatás

1. Objective. Measurement of high speed signals requires a measurement device with large bandwidth. The demands against the measurement system for calibration of high speed pulse and impulse generators are very high. The bandwidth of the measurement device cannot be increased beyond a certain limit. The solution for extending the bandwidth is to postprocess the measured signals and correct the results numerically. The Electronic Instrumentation and Metrology Group, Electricity Division, NIST, calibrates high-speed pulse and impulse generators, providing the national standard for these measurements. To reach the very high bandwidth the calibration system incorporates an inverse filtering algorithm (deconvolution) to improve the measurement by means of digital postprocessing of the measured signal. Postprocessing (reconstruction) of the measured signal is an estimation procedure, since the measurement is always corrupted by noise. The reconstructed signal will differ from the true input signal. There is an increased interest for the error analysis of the reconstructed signal. An error bound around the reconstructed waveform would give an idea about the possibilities of the true input signal. We can distinguish between systematic and stochastic errors. The primary aim of this research is to estimate the systematic error (bias) of the reconstruction process. 2. Expected benefits. The error analysis of the signal reconstruction gives a bound around the estimated signal. This information increases the usefulness of the reconstruction with a great extent, since the uncertainty and bias envelope provides an upper bound to the errors with a certain confidence level.

© 2010-2017 BME MIT