Fusion of kinetic data using composite filters

This paper describes a single method for improving the estimation of kinetic data, i.e. acceleration, velocity or displacement, from two (or all three) raw data measurements, via the use of ‘composite filters’. The improvements are intended to provide estimates of unmeasured data (e.g. velocity), and to obtain more consistent estimates of measured data over a wider frequency range. The method makes use of complementary filters 1 for the synthesis of the composite filter, and experimental results are included in the paper to emphasize the effectiveness of the technique. Comments are also included on the use of the method for the synthesis of integrated demand signal data with zero bias or drift.