ESTIMATION OF SATELLITE ALTIMETER ECHO-SIGNAL PARAMETERS BY STATISTICAL FITTING METHODS IN THE COURSE OF RETRACKING

• Published in: Izvestiâ vysših učebnyh zavedenij Rossii. Radioèlektronika no. 1; pp. 5 - 16
• Main Authors: Borovitsky, D. S., Zhesterev, A. E., Ipatov, V. P., Mamchur, R. M.
• Format: Journal Article
• Language: English; Russian
• Published: Saint Petersburg Electrotechnical University "LETI" 28.02.2019
• Subjects: дискриминатор; дообработка; метод наименьших квадратов; оценка по максимуму правдоподобия; подгонка; спутниковый высотомер; физическая модель; эхосигнал
• ISSN: 1993-8985; 2658-4794
• DOI: 10.32603/1993-8985-2019-22-1-5-16

Satellite radar altimeters play a key role in numerous space missions for the remote Earth sensing. The data they provide are used in solving various fundamental and applied problems of geophysics, oceanography, meteorology, etc. In many modern measuring systems the altimeter data is processed in several stages. One of them is the ground-based retracking of the information streamed from the spacecraft. The goal of this work is to study altimeter parameter estimators in the course of the satellite altimeter data retracking by ground-based complex. The main task of delay-lock loop onboard a satellite carrier is a reliable keeping of received echo-signal within the tracking window. More accurate estimates of information parameters are worked out by the ground segment where data from the satellite is delivered via telemetry. Retracking can be performed either without using any physical echo model, or relying on some mechanism under generation of response of an illuminated surface to the probing signal. In the latter case, the measuring results become more trustworthy. The paper deliberations are based on the model close to the classical Brown’s one, they include algorithms of its parameters statistical fitting to the observation and computer simulation of fitting according to the maximal likelihood (ML) and the least squares (LS) methods. The results obtained are compared to the potential attainable and show that while LS fitting yields noticeably to the potential, experimental accuracy of ML-fitting practically coincides with the potential one.