APPLICATION OF PREDICTION-BASED PARTICLE FILTERS FOR TELEOPERATIONS OVER THE INTERNET

摘要

In this paper, we present the prediction-based particle filterrnapproach for processing motion and force data inrnteleoperation over the Internet. We first introduce thernprediction-based particle filter algorithm, one of thernSequential Monte Carlo methods based on the recursivernBayesian prediction. The prediction algorithm is applied torndynamic models of the motion and force data flows in thernstate-space formulation. It is applied to the motion datarntransmitted to the slave controller and to the reflecting forcerndata received at the master controller. Experiments arernperformed using the haptic device within a virtual 3Drngraphical environment. In each experiment, the motion andrnreflecting force data extracted from the haptic device are usedrnto verify the prediction performance of the proposed method.