Determining Kalman Filter Input Noises Using PES Pareto

One of the key questions that befalls anyone about to design a Kalman Filter is the question of determining the noise inputs. As methodical and systematic as Kalman filter design is, it is wholly dependent upon the choice of noise source values, and these are usually ad-hoc. On the other hand, the PES Pareto Methodology [1] provides a systemic method for modeling and measuring broadband noises in a feedback loop. As with the old Reece's Cup commercials, it seems that one might want to know if these two techniques can be combined, so that we obtain well quantified process and sensor noise estimates based upon our PES Pareto based measurements. That is the subject of this paper.