Title :
A single-chip QR decomposition processor for extended square-root Kalman filters
Author_Institution :
Advanced Micro Devices Inc., Sunnyvale, CA, USA
Abstract :
A single-chip architecture of a dedicated-hardware QR decomposition processor (QRDP) for the extended square-root Kalman filter is presented. As a slave peripheral to a 32 bit host processor, this chip acts as a numeric coprocessor which exclusively performs the QR decompositions for measurement and time updates by using the sequences of fast Givens rotations and elementary matrices. QRDP can be programmed to implement the square-root Kalman filter with various dimensions. The maximum allowable numbers of measurements, process states, and noises are 6, 12, and 12, respectively. The application of QRDP considered is target-tracking Kalman filters. For the nine-state case, a simulation study shows that QRDP can achieve sub-millisecond performance for both QR decompositions for measurement and time updates
Keywords :
Kalman filters; digital filters; digital signal processing chips; digital filters; elementary matrices; extended square-root Kalman filters; fast Givens rotations; measurement updates; numeric coprocessor; single-chip QR decomposition processor; target-tracking; time updates; Computer architecture; Coprocessors; Covariance matrix; Hardware; Matrix decomposition; Noise measurement; Performance evaluation; Semiconductor device measurement; State estimation; Time measurement;
Conference_Titel :
Signals, Systems and Computers, 1991. 1991 Conference Record of the Twenty-Fifth Asilomar Conference on
Conference_Location :
Pacific Grove, CA
Print_ISBN :
0-8186-2470-1
DOI :
10.1109/ACSSC.1991.186504