Controlled mechanical systems with friction

摘要: Grey-box modeling covers the domain where we want to use a balanced amount of white-box based on first principles and black-box empiricism. The two grey-box models presented combine model with model, i.c., Neural Network Polytopic that are capable identifying friction characteristics left unexplained by modeling. In an experimental case-study, both applied identify rotating arm subjected friction. An augmented state extended Kalman filter is used iteratively off-line for estimation unknown parameters. For studied example defined topologies, little difference observed between models. addition, applicability identified illustrated in compensation control scheme objective linearize system. This chapter has been published European Journal Control [62]. Extended Filter tool this documented more detail [61].