Constraint-based Approaches for Robotic Systems: from Computer Vision to Real-Time Robot Control

Abstract

This thesis presents a framework to model and solve diverse problems using constraint-based approaches. The primary contribution is an intelligent robotic system composed of four classical problems formulated using constraints: robot motion planning, real-time robot motion control, object detection and object tracking. Our results are used for the development of intuitive robot programming interfaces and in applications such as manufacturing and logistics.