Title :
Predictive transform: further results
Author_Institution :
Dept. of Appl. Sci., Coll. of Staten Island, City Univ. of New York, NY, USA
Abstract :
It is shown that statistics used to design a minimum mean square error (MSE) linear predictive transform (LPT) estimator result in significant design and implementation simplifications, by factors of eight and four, respectively, when these statistics satisfy certain symmetry constraints and the state dimensionality is large. More specifically, the computational burden associated with the design of the exact minimum MSE LPT signal source model present in the LPT estimator is reduced by a factor of approximately eight; the computational burden linked to the design of LPT estimator gains is simplified by a factor of approximately eight; and the implementation burden of the LPT estimator is reduced by a factor of approximately four
Keywords :
computerised signal processing; estimation theory; statistics; transforms; computerised signal processing; linear predictive transform estimator; minimum mean square error; state dimensionality; statistics; symmetry; Covariance matrix; Decoding; Educational institutions; Error analysis; Kalman filters; Lagrangian functions; Mean square error methods; Signal design; State estimation; Statistics;
Conference_Titel :
Aerospace and Electronics Conference, 1990. NAECON 1990., Proceedings of the IEEE 1990 National
Conference_Location :
Dayton, OH
DOI :
10.1109/NAECON.1990.112750