Constrained Robotic Control - Final Project - Constrained Task-Space Control Using CBF–QP
Автор: Cayo Almeida
Загружено: 2025-12-10
Просмотров: 13
Описание:
This video presents the development and simulation of a constrained task-space control system for a UR5 manipulator, designed to ensure safe and reliable execution of manipulation tasks under geometric and kinematic constraints.
The project is part of the Final Practical Assignment for the Constrained Robotic Control course and aims to apply advanced concepts of control, safety enforcement, and multi-stage task execution within a simulated environment.
The video demonstrates the main components of the proposed solution:
• Task-space tracking using an analytical geometric Jacobian and a resolved-motion control law;
• Real-time safety enforcement through Control Barrier Functions (CBFs), preventing violations of joint limits, workspace boundaries, and self-collision distances;
• Integration of CBFs into a Quadratic Program (QP), ensuring minimal deviation from the nominal motion while maintaining full constraint satisfaction;
• Definition of geometric primitives for continuous distance monitoring between robot links and workspace surfaces;
• Execution of a complete pick-and-place routine, coordinated by a finite-state machine that governs approach, descent, grasping, lifting, transport, release, and retraction stages;
• Handling of infeasible targets, where the CBF–QP controller properly rejects motions that would lead the robot into unsafe configurations.
The video presents the full sequence of the experiment—from the initial configuration to the completed manipulation attempts—highlighting how the controller integrates task tracking, safety constraints, and behavioral supervision within the simulated environment.
ABOUT THE PROJECT
This work was developed for the Constrained Robotic Control course at the Federal University of Minas Gerais (UFMG).
Student: Cayo Fábio de Almeida Sousa
Professor: Vinícius Mariano Gonçalves
Term: 2nd Semester of 2025
⭐ TECHNOLOGIES USED
Simulator: CoppeliaSim (with ZMQ Remote API)
Programming Language: Python
Manipulator: UR5
Control Method: CBF–QP (Control Barrier Functions + Quadratic Programming)
Повторяем попытку...
Доступные форматы для скачивания:
Скачать видео
-
Информация по загрузке: