Learning Feedforward Control Using Multiagent Control Approach for Motion Control Systems

Multiagent control system (MACS) has become a promising solution for solving complex problems. Using the advantages of MACS-based design approaches, novel advanced mechatronic systems been developed in this paper. The study aimed at integrating learning into MACS. Specifically, feedforward (LFFC) is implemented as pattern incorporation major novelty work that feedback part realized real-time periodic MACS, while LFFC algorithm done on-line, asynchronously, and separate non-real-time aperiodic As result, method developed. A second-order B-spline neural network (BSN) used function approximator whose input-output mapping can be adapted during intended to equal inverse model plant. To provide features system, open robot software (OROCOS) employed development runtime environment. case using simulated linear motor presence nonlinear cogging friction force well mass variations illustrate proposed method. designed consists setpoint generator, controller, time-index learn on-line. Simulation results have demonstrated applicability