• 9 months at INRIA Sophia-Antipolis. • 9 months at INRIA Rhône-Alpes
Key Words : Robotics, Navigation systems, Perception, Prediction, Human aware navigation, Assistance to frail people.
Summary : This postDoc takes place in the scope of the Inria Large Scale Initiative Action AEN PAL (Personally Assisted Living http://pal.inria.fr/) which gathers the research activities of several Inria research teams to improve the autonomy and the quality of life for the elderly and fragile persons. The Action aims to create a research infrastructure that will offer technology solutions to improve the quality of life of frail people and their proximity circle. It focuses on the autonomy of the elderly in home health care by securing, preserving or restoring their functions necessary for everyday life. Subject : The objective of this postdoc is to develop navigation technologies adapted to the mobility assistance of aging or disabled people. The core of this work relies on complementary skills of Arolag and e-Motion INRIA projects concerning mapping, localization and safe navigation in dynamic and human populated environments.
Autonomous robotic systems are progressively moving from controlled laboratory environments towards real-world applications. State of the art approaches now endow robots with the ability to navigate within complex real-world environments by taking into account both the robot’s surroundings and its position within the environment. Additionally, the perceptual capabilities of these systems have recently been augmented through the use of realtime visual sensors that provide a very rich source of information [Meilland10]. One of the major problems remains the capability to deal with large variations of the scene over time. A first objective will be to develop a novel approach to capture and characterize the information shared by different views of a dynamic and evolving environment registered under different conditions of observations. Doing this, we expect to get a representation well adapted to autonomous navigation tasks and robust to the changes in the environment. Relying on this representation, a second objective will be to develop navigation strategies that allow the robot to safely move in changing environment, with a socially acceptable behaviour. In the scope of the postDoc, we propose to integrate previous works developped both in Arolag and e-Motion research teams adressing the navigation decision process [Fulgenz10], the spherical robot-centered representation of the environment described in [Meilland11], abilities to predict the near future of this perceived environment [MVAVasq 2007], and social conventions [IROSRios 2011]. The core of the work will be focused on dynamic and evolving environments. By dealing with dynamic environment, we think of objects present in the environment moving in the scene (for example people) during the robot navigation. By evolving environment, we consider that configurations of the scene can change during time (door open or closed, chair, medical devices…)
Robustness, adaptability and cost of implementation of the developed technologies will be a major requirement.
The approaches will be implemented on two experimental platforms : an automated wheelchair (INRIA platform at Grenoble) and a smart walking aid (INRIA platform at Sophia-Antipolis).
[Fulgenz10] Fulgenzi C ; Spalanzani A ; Laugier C ; Tay C. “Risk based motion planning and navigation in uncertain dynamic environment” Rapport de recherche INRIA, 14 pages, Octobre 2010.
[IROSRios 2011] Rios-Martinez J. , Spalanzani A., Laugier C., "Probabilistic autonomous navigation using Risk-RRT approach and models of human interaction", in : Proceedings of the 2011 IEEE/RSJ International Conference on Intelligent Robots and Systems, San Francisco, USA sept. 2011.
[Meilland10] Meilland, M., Comport, A.I. and Rives,P. "A Spherical Robot-Centered Representation for Urban Navigation ", IEEE/RSJ International Conference on Intelligent Robots and Systems, Taipei, Taiwan, October 2010
[Meilland 11] Dense visual mapping of large scale environments for real-time localisation. In IEEE/RSJ International Conference on Intelligent Robots and Systems. San Francisco, California, USA, Sept. 2011
[MVAVasq 2007] Dizan Alejandro Vasquez Govea, Thierry Fraichard, Christian Laugier, “Intentional Motion On-line Learning and Prediction” - Machine Vision and Applications 2007
Requirements For this position, candidates with a PhD focusing on computer vision, robotics, visual servoing, and/or automatic control, will be considered. The candidate should have also good abilities for software development (C++, linux, real time algorithms) and experimental issues.
Send a CV and a motivation letter