N. A. Khovanova

Artificial neural networks in hard tissue engineering: Another look at age-dependence of trabecular bone properties in osteoarthritis

Artificial Neural Network (ANN) model has been developed to correlate age of severely osteoarthritic male and female specimens with key mechanical and structural characteristics of their trabecular bone. The complex interdependency between age, gender, compressive strength, porosity, morphology and level of interconnectivity was analysed in multi-dimensional space using a two-layer feedforward ANN. Trained by Levenberg-Marquardt back propagation algorithm, the ANN achieved regression factor of R = 96.3% between the predicted and target age when optimised for the experimental dataset. Results indicate a strong correlation of the 5-dimensional vector of physical properties of the bone with the age of the specimens. The inverse problem of estimating compressive strength as the key bone fracture risk was also investigated. The outcomes yield correlation between predicted and target compressive strength with the regression factor of R = 97.4%. Within the limitations of the input data set, the ANNs provide robust predictive models for hard tissue engineering decision support.

Generalised stochastic model for characterisation of subcutaneous glucose time series

A generalised stochastic model with second order differential equations is proposed to describe the response of blood glucose concentration to meals in groups of nondiabetic people and two types of diabetic patients. A variational Bayesian approach is applied in order to infer parameters of the models, and the best model was selected based on the computed log-evidence for each prandial event. The model with a linear structure represents most of the events, while the nonlinear terms need to be included more frequently for Type II diabetic patients. This indicates different physiological mechanisms of glucose absorption for different groups. The deterministic parameters and intensities of stochastic components are compared by groups using the ANOVA test, and the results show significant differences between the groups. This model can potentially be used for long term prediction of the glucose concentration response to external stimuli.

Frequency response of an energy harvester to harmonic noise: Towards stochastic frequency response of nonlinear systems

Energy harvesting of ambient energy of different forms has attracted considerable attention of researchers from both theoretical and applied fields. Whereas a typical design of energy harvester is based on a resonant structure — a linear oscillator — aimed at a harmonic excitation, in practice the behaviour of a harvester deviates from linear, and ambient excitations have a much more complex structure than a simple sine-wave. Therefore, it is important to develop a framework for a nonlinear harvester design that would target the stochastic nature of vibrations for effective energy harvesting. In this paper such framework is based on stochastic frequency response. Using the suggested framework the performance of a harvester model for soft and hard nonlinearities in stiffness was compared with the linear harvester. The comparison indicates the presence of a novel phenomena linked to stochastic bifurcations observed by varying characteristics of harmonic noise.