Optimisation of regularisation methods for differentiation of measurement data in monitoring of human movements

The research reported in this paper is related to the regularised differentiation of measurement data from impulse-radar sensors and infrared depth sensors applied in systems for unobtrusive monitoring of elderly persons. Four strategies for selection of regularisation parameters are compared in terms of their potential of decreasing the uncertainty of walking velocity estimation. The comparison is based on synthetic and real-world data. The best results are obtained by means of a strategy based on the so-called Stein's unbiased risk estimator.