A more general incremental inter-agent learning adaptive control for multiple identical processes in mass production

Abstract

Mass production has been a trend for modern manufacturing. Information from peer processes may be used to improve the individual process control performance. The idea of ‘incremental step mimicking’ has been presented in our previous work [1] which resulted in ‘incremental inter-agent learning’ (IIAL) adaptive control. However, that work is based on a primitive adaptive control in which the control input is directly calculated from the online-identified model. This paper explores the inter-agent learning adaptive control based on a more complicated adaptive control, and proposes a more general Full-Scale IIAL (FS-IIAL) of which the previous IIAL [1] can be viewed as a special case. With the ensured robust stability, the proposed work is applied on the case when LIP formulae is a single layer RBF neural network, and simulation result validates the superior control performance of each process over the original adaptive control and the previous IIAL.