Title :
Analyzing bilinear neural networks with new curve fitting for application to human motion analysis
Author :
Ito, Toshio ; Senta, Yosuke ; Nagashima, Fumio
Author_Institution :
Human Centric Comput. Labs., Fujitsu Labs. Ltd., Atsugi, Japan
Abstract :
This paper presents a method for the analysis of bilinear neural networks (BNNs) with time delay. A BNN system was proposed to analyze a weak nonlinear model. To analyze this system, we need to solve the equations for the system. In this paper, we propose a method that determines whether it is possible to solve the equations analytically. If this is possible, we propose a concrete method for solving them. This paper also presents new optimum curve fitting for data sampled at discrete times, such as data outputted from a sensor in a mobile phone, and a new BNN system for curve fitting. Moreover, this paper presents one example where the BNN system for curve fitting and other BNN systems are applied to the analysis of human motion.
Keywords :
behavioural sciences computing; curve fitting; data analysis; neural nets; BNN system; bilinear neural network; curve fitting; human motion analysis; mobile phone; nonlinear model; time delay; weak nonlinear model analysis; Curve fitting; Equations; Finite impulse response filter; Low pass filters; Mathematical model; Neurons; Neural network; accelerometer; analytic solution; curve fitting; human motion analysis; mobile phone; nonlinear model;
Conference_Titel :
Systems, Man, and Cybernetics (SMC), 2012 IEEE International Conference on
Conference_Location :
Seoul
Print_ISBN :
978-1-4673-1713-9
Electronic_ISBN :
978-1-4673-1712-2
DOI :
10.1109/ICSMC.2012.6377725
