Nonconvex Activation Noise-Suppressing Neural Network for Time-Varying Quadratic Programming: Application to Omnidirectional Mobile Manipulator

Abstract

This paper proposes an improved general zeroing neural network (ZNN) model to suppress noise and to enhance the real-time performance of solving time-varying quadratic programming (TVQP) problems. The proposed model allows nonconvex activation functions (AFs) and has noise suppression characteristics, i.e., the nonconvex constrained noise suppressed ZNN (NCNSZNN) model. Theoretical analyses show that the developed NCNSZNN model converges globally to an accurate solution to the TVQP problem and is robust in the case of measurement noise (MN). Illustrative examples and comparisons are supplied to verify the validity and superiority of the proposed model for online solving TVQP constrained by equalities and inequalities (EAI) with MN.