High-Resolution Touch Floor System Using Particle Swarm Optimization Neural Network

Abstract

A touch floor system, based on force sensitive resistors, capable of identifying user position and motion with high resolution, is proposed in this paper. A particle swarm optimization-based neural network (NN), initialized with the output of a Levenberg-Marquardt-based NN, allows inaccuracy drawbacks of the trilateration method in position estimation due to sensor's nonlinearity to be reduced to one fifth under non-stationary conditions. Furthermore, position-tracking accuracy is improved by a Kalman filter and a motion recognition algorithm is suggested for mimicking computer mouse clicks. Experimental results show non-uniformly sized icons displayed with high-resolution coordinates can be selected on the floor by the participants of diversified weights. This proves the feasibility of a high-resolution touch floor interface scalable for large area, by facilitating digitally mediated human-architecture interactions.