Use of Sequential Differences in Smoothing 3-D Data
Abstract
The use of sequential (successive) differences in detecting wild data or outliers in data collected on torpedo paths is explored. It is assumed that the collected data on torpedo paths is of the form: x(i) = x(t(i)) = P(x)(t(i)) + n(i) + d(i) where P(x)(t(i)) is the x components of the position of the torpedo at time t with P(t) being some appropriate low order polynomial, n(i) is the 'noise' due to measurement error, and d(i) is a perturbation or disturbance which if present in sufficient amplitude will cause x(i) to be a 'wild' datum or outlier. Variations in the patterns (signatures) of successive differences caused by a variety of perturbations are examined for the purpose of setting thresholds to be used to detect outliers. Data collected from a torpedo path at NGWES are used for illustration.
Document Details
- Document Type
- Technical Report
- Publication Date
- May 01, 1979
- Accession Number
- ADA522451
Entities
People
- J. B. Tysver
Organizations
- Naval Postgraduate School