Fast orthogonal search for array processing and spectrum estimation

Speed Sensorless Estimation of AC Induction Motors Using the Fast Orthogonal Search Algorithm

“…The fast orthogonal search (FOS) algorithm [8], [9], [13], [14] is a nonlinear technique that allows fitting of a time series y(n), of length N, to a functional expansion of arbitrary functions p m (n). The functional expansion is given by…”

“…The fast orthogonal search (FOS) algorithm has been shown to have a better (smaller) frequency resolution than the FFT [6]- [9]. Thus, FOS can maintain the same frequency resolution as the FFT, while using a shorter time series.…”

Interpreting Microarray Data and Related Applications Using Nonlinear System Identification

“…However, the latter method has computational complexity and storage requirement dependent upon the square of the number of candidate terms that are searched, while in FOS the dependence is reduced to a linear relationship. In addition FOS and/or iterative forms [44 -47] of FOS have been used for high-resolution spectral analysis [42,45,47,48], direction finding [44,45], constructing generalized single-layer networks [46], and design of two-dimensional filters [49], among many applications. Wu et al [50] have compared FOS with canonical variate analysis for biological applications.…”

Accurate Resolution of Signals Using Integer-Coded Genetic Algorithms

“…It was also applied to model clean and noisy time series data using non-Fourier sinusoidal series representation [2]. An iterative version of FOS (IFOS) [6] was also developed and applied to the problem of estimating the temporal frequencies of narrow-band signals in noise. Chon [7] employed the FOS algorithm to detect sinusoidal components buried in noise and observed in physiological signals such as heart rate and renal blood pressure and flow data.…”

“…(phone: +2 (02) 413-5457; fax: +2 (02) 418-6945; email: h.abbas@ieee.org.) minimum in the MSE as a function of the parameter estimates [6]. Evolutionary methods [8], [9], on the other hand, are population based algorithms which are modeled loosely on the principles of the evolution via natural selection.…”