"A number of outstanding problems in robotic motion and manipulation involve tasks where degrees of freedom (DoF), be they part of the robot, an object being manipulated, or the surrounding environment, cannot be accurately controlled by the actuators of the robot alone. Rather, they are also controlled by physical properties or interactions - contact, robot dynamics, actuator behavior - that are influenced by the actuators of the robot. In particular, we focus on two important areas of poorly controlled robotic manipulation: motion planning for deformable objects and in deformable environments; and manipulation with uncertainty. Many everyday tasks we wish robots to perform, such as cooking and cleaning, require the robot to manipulate deformable objects. The limitations of real robotic actuators and sensors result in uncertainty that we must address to reliably perform fine manipulation. Notably, both areas share a common principle: contact, which is usually prohibited in motion planners, is not only sometimes unavoidable, but often necessary to accurately complete the task at hand. We make four contributions that enable robot manipulation in these poorly controlled tasks: First, an efficient discretized representation of elastic deformable objects and cost function that assess a ``cost of deformation' for a specific configuration of a deformable object that enables deformable object manipulation tasks to be performed without physical simulation. Second, a method using active learning and inverse-optimal control to build these discretized representations from expert demonstrations. Third, a motion planner and policy-based execution approach to manipulation with uncertainty which incorporates contact with the environment and compliance of the robot to generate motion policies which are then adapted during execution to reflect actual robot behavior. Fourth, work towards the development of an efficient path quality metric for paths executed with actuation uncertainty that can be used inside a motion planner or trajectory optimizer."

Planifier les horaires des trains qui circulent sur un réseau constitue une tâche complexe. En Europe, elle s'inscrit dans un processus qui voit le gestionnaire d'infrastructure opérer des choix pour répartir une ressource rare, la capacité d'infrastructure, entre des usages concurrents - commerciaux (trains de voyageurs et de marchandises) et de maintenance (travaux) - et des demandeurs multiples. Dans ce travail, nous cherchons à mieux cerner comment s'organise et fonctionne le processus pour le réseau ferré français par le prisme du fret, activité ouverte à la concurrence depuis 2006 mais qui traverse une crise structurelle depuis plusieurs décennies. Les différentes facettes temporelles du processus de répartition des capacités constituent le fil conducteur de notre analyse. A travers une démarche croisant le point de vue du gestionnaire d'infrastructure et celui de ses clients, au premier rang desquels les entreprises ferroviaires, nous montrons que le fret occupe une place singulière dans ce processus qui se structure selon un calendrier spécifique, fruit de dispositions règlementaires européennes et d'héritages et de choix nationaux. La question de la qualité de sillons offerts par le gestionnaire d'infrastructure est appréhendée sous l'angle d'une articulation entre composantes de court et de long termes par rapport à l'échéance de circulation. Activité plurielle, le fret a besoin de flexibilité et de visibilité. Nous nous intéressons enfin aux sillons-catalogue en tant que dispositif participant à garantir de la capacité au fret. Une analyse de la méthode allemande, à travers le projet neXt, est proposée en guise de contrepoint. Nous mettons en évidence que l'équilibre entre des exigences industrielles (standardisation) et commerciales (personnalisation) dans la production des sillons est indispensable mais demeure délicat à établir / Train timetabling on a rail network is a complex task. In Europe, this issue lies within a process for which the infrastructure manager has to consider competing uses - passenger and freight trains as well as maintenance work - and several users of a scarce resource, infrastructure capacity. The aim of this research is to better understand the way this process is organised and works on the French rail network. We have chosen to tackle this issue focusing on freight as an activity which has been opened up to competition since 2006 but has experienced structural difficulties for decades. The various time aspects of the capacity allocation process are the key focus of our analysis. Our approach considers the perspectives of the infrastructure manager and of their customers, foremost among which are train operating companies. We highlight that freight occupies a unique place in this process which is based on a schedule resulting from both European regulations and the legacy of decisions taken at the national level. The issue of the quality of the train paths offered by the infrastructure manager is studied with a focus on the links between short and long term requirements with regard to the train running deadline. In spite of their diversity, we demonstrate that freight operators need flexibility and visibility. Finally, we analyse catalogue paths as a specific offer that aims to guarantee capacity for freight. The German method, through the neXt project, provides an alternative insight into this issue. Accommodating freight train paths into the timetable is essentially a necessary but delicate balance of interests: standardisation vs. customisation