Scaling and Roundoff in Fixed-Point Implementation of Control Algorithms

Abstract
The analysis and design of digital controllers on microprocessors using fixed-point arithmetic have been considered. Three scaling techniques are applied to the design of digital controllers in such a way as to avoid overflow and to optimize the signal-to-roundoff noise ratio. Both the scaler and the multivariable cases are treated. Statistical estimates of the steady-state roundoff and quantization errors are derived.

This publication has 5 references indexed in Scilit: