Reactive Solutions to Optimal Motion Planning for Mobile Robots using Reinforcement Learning

Rousseas, Panagiotis

dc.contributor.author	Rousseas, Panagiotis
dc.date.accessioned	2025-09-22T09:14:27Z
dc.date.available	2025-09-22T09:14:27Z
dc.identifier.uri	https://dspace.lib.ntua.gr/xmlui/handle/123456789/62482
dc.identifier.uri	http://dx.doi.org/10.26240/heal.ntua.30178
dc.rights	Αναφορά Δημιουργού-Μη Εμπορική Χρήση-Όχι Παράγωγα Έργα 3.0 Ελλάδα	*
dc.rights.uri	http://creativecommons.org/licenses/by-nc-nd/3.0/gr/	*
dc.subject	Optimal Control	en
dc.subject	Safe Control	en
dc.subject	Motion Planning	en
dc.subject	Robust Control	en
dc.subject	Reinforcement Learning	en
dc.subject	Βέλτιστος Έλεγχος	el
dc.subject	Ασφαλής Έλεγχος	el
dc.subject	Σχεδιασμός Πορείας	el
dc.subject	Εύρωστος Έλεγχος	el
dc.subject	Ενισχυτική Μάθηση	el
dc.title	Reactive Solutions to Optimal Motion Planning for Mobile Robots using Reinforcement Learning	en
dc.contributor.department	Control Systems Laboratory	el
heal.type	doctoralThesis
heal.classification	Robotics	en
heal.classification	Control Theory	en
heal.language	el
heal.language	en
heal.access	free
heal.recordProvider	ntua	el
heal.publicationDate	2025-03-12
heal.abstract	Motion planning is one of the most fundamental problems of Robotics and has been a focal point for research since the infancy of the field. Autonomous operation and task accomplishment requires robotic navigation in the real world, where obstacles may pose a threat to the robot’s integrity and require careful examination to ensure safety. At the same time, modeling uncertainty, sensor noise and unknown obstacles require robust solutions and the integration of high-level planning (e.g. finding valid paths within the workspace) with low-level controllers, which handle a robot’s dynamics. To this end, this Thesis concentrates on solutions to various aspects of optimal motion planning through underlying position-feedback velocity fields. Treating problems from static/dynamic workspaces to disturbances and higher-order models, the underlying velocity fields are demonstrated to provide several benefits in real-world robotic control with an emphasis on mathematical rigor. The aim is to bridge the gap between high and low-level control without sacrificing provable guarantees of safety, convergence and –to the degree that it is possible– optimality.	en
heal.sponsor	The research work was supported by the Hellenic Foundation for Research and Innovation (HFRI) under the 4th Call for HFRI PhD Fellowships (Fellowship Number: 9110).	en
heal.advisorName	Κυριακόπουλος, Κωνσταντίνος
heal.committeeMemberName	Κυριακόπουλος, Κωνσταντίνος
heal.committeeMemberName	Ευάγγελος, Παπαδόπουλος
heal.committeeMemberName	Δημαρόγκωνας, Δήμος
heal.committeeMemberName	Μπεχλιούλης, Χαράλαμπος
heal.committeeMemberName	Πουλακάκης, Ιωάννης
heal.committeeMemberName	Τζαφέστας, Κωνσταντίνος
heal.committeeMemberName	Κορδώνης, Ιωάννης
heal.academicPublisher	Σχολή Μηχανολόγων Μηχανικών	el
heal.academicPublisherID	ntua
heal.numberOfPages	323
heal.fullTextAvailability	false