ELSS-DZN and ELSS-IFDHGZN Algorithms Solving Future Quadratic Programming Applied to Robot Manipulator

Pengfei Guo,Yunong Zhang

DOI: https://doi.org/10.1109/tsmc.2023.3331051

2024-01-01

IEEE Transactions on Systems Man and Cybernetics Systems

Abstract:It is common knowledge that the future quadratic programming (FQP) problem is a challenging and widely applicable topic in mathematical science and many engineering fields. In this article, we systematically derive a new specific explicit linear left and right six-step (ELSS) rule with order-7 precision. With the help of the ELSS rule, the ELSS-type discretized Zhang neurodynamics (ELSS-DZNs) algorithm that develops from continuous Zhang neurodynamics (CZNs) model is proposed for tackling the FQP problem subject to linear equation (LE), while the inverse-free continuous hybrid gradient Zhang neurodynamics (IFCHGZNs) model, which combines the Getz–Marsden dynamic system (GMDS) model 2 for time-variant inverse computing (TVIC) and CZN model for tackling the time-variant QP (TVQP) problem subject to LE, develops the ELSS-type inverse-free discretized hybrid gradient Zhang neurodynamics (ELSS-IFDHGZN) algorithm for dealing with the FQP problem subject to LE. Theoretically, three theorems are presented to show the convergent properties of the proposed general CZN and IFCHGZN models by Lyapunov stability theory, while three theorems and five corollaries are presented to show the truncation error pattern of the ELSS-DZN and ELSS-IFDHGZN algorithms by using the stability and convergence theory of linear multistep formulas. Moreover, two numerical experiments are adopted to demonstrate the availability and convergent properties of the proposed general ELSS-DZN and ELSS-IFDHGZN algorithms. In the end, numerical and physical experiments based on two kinds of robot manipulators are additionally performed to verify the applicability and validity of the proposed general ELSS-DZN and ELSS-IFDHGZN algorithms.