Error propagation in the recursive QRD LS filter

Abstract
The signal processing interpretation of the state variables of the recursive QR decomposition LS (least-squares) algorithm is analyzed. This makes it possible to find bounds for the state variables in order to ensure a correct fixed-point implementation. Furthermore, a general model of the propagation of quantization errors has been investigated; this model exhibits the property of exponential stability if a single error is inserted at a fixed time.

This publication has 4 references indexed in Scilit: