Modelling haptic devices using a rule-based expert system

Modelling physical systems is one of the main technical challenges in the field of control systems and haptic devices. In this work, a two part methodology is proposed that generates a model which employs qualitative reasoning to encapsulate nonlinear effects that are often approximated as linear processes. We utilize fuzzy set theory to implement a rule-base that has been constructed from both conventional and expert knowledge to model the nonlinear damping behaviour of the PHANToM/spl trade/ haptic device. Our methodology is used to produce an estimate of the nonlinear parameters in a mathematical model a PHANToM/spl trade/ haptic device integrated with a rate feedback controller. Most estimation methods approximate damping factors using a linear approximation based on experimental data. In the first part of our method, a rule-based expert system is developed based on constant parameters which have been estimated from experimental data (linear model) and the system parameters are tuned for multiple operating regions. The second part of our method develops an expert system using constant parameters based on expert knowledge. Several system responses are examined to show the ability of our technique to capture a variety of conditions.