کنفرانس مهندسی برق ایران

صفحه اصلی / سی و چهارمین کنفرانس بین المللی مهندسی برق

Multilateral Nonlinear Model Predictive Control for Telerehabilitation with Serial Manipulators

نویسندگان :

Alirezaz Gholami¹ Iman Sharifi²

1- Amirkabir University of Technology 2- Amirkabir University of Technology

کلمات کلیدی :

Nonlinear Model Predictive Control،Feedback Linearization،Multi-Agent System،Multilateral Teleoperation،Graph Laplacian،Decentralized Coordination،Degrees of Freedom،Integral of Timeweighted Absolute Error،Serial Manipulators

چکیده :

This paper presents a Nonlinear Model Predictive Control (NMPC) scheme for a two Degrees of Freedom (DOF) serial manipulator operating within a Multilateral Teleoperation (MT) network. The plant is regulated via Feedback Linearization (FL), and the NMPC solves a finite-horizon optimization using a time-invariant nominal model and weights. A Multi-Agent System (MAS) formulation with a connected Graph Laplacian (GL) coordinates the network; Decentralized Coordination (DC) through local damping improves transients and follows established MASbased teleoperation designs intended to preserve synchronization under packet drops and time-varying delays. Controller tuning is clarified by mapping the closed loop to a standard secondorder template that links natural frequency, damping ratio, prediction horizon, and weighting factors. Simulation studies, including a mirror-therapy scenario and a “maple-leaf” trajectory, show accurate tracking, smooth torques, and predictable Integral of Time-weighted Absolute Error (ITAE) / effort trade-offs as communication bandwidth increases. Bench-top experiments on a low-cost embedded stack corroborate deployability.