Section: New Results

Visual navigation of mobile robots

Automous navigation of wheelchairs

Participants : Vishnu Karakkat Narayanan, François Pasteau, Marie Babel, François Chaumette.

The goal of this work is to design an autonomous navigation framework of a wheelchair by means of a single camera and visual servoing. We first focused on a corridor following task where no prior knowledge of the environment is required. The servoing process matches the non-holonomic constraints of the wheelchair and relies on two visual features, namely the vanishing point location and the orientation of the median line formed by the straight lines related to the bottom of the walls. This overcomes the initialization issue typically raised in the literature [9] . The control scheme has been implemented onto a robotized wheelchair and results show that it can follow a corridor with an accuracy of ±3 cm [16] .

We then focused on a door passing task [33] . This doorway passing (and corridor turning) task employs the position of a single doorpost in the image as an input to a Lyapunov-based control scheme which allows the wheelchair to take up a desired trajectory about the doorpost. This trajectory avoids collision with the wall and guarantees that the wheelchair positions itself in front of the doorway regardless of its initial position. Results in simulation demonstrate the convergence and robustness of both control schemes. Experiments conducted on a wheelchair indicate the validity of applying the proposed low-level control system [16] .

This study is in the scope of the Inria large-scale initiative action PAL (see Section  8.2.6 ) as well as of the Apash project (see Section  8.1.1 ).

Semi-autonomous control of a wheelchair for navigation assistance along corridors

Participants : Vishnu Karakkat Narayanan, Marie Babel, François Pasteau, Alexandre Krupa.

This study concerns a semi-autonomous control approach that we designed for safe wheelchair navigation. The control relies on the combination of primary tasks of wall avoidance as well as door passing performed by a dedicated visual servoing framework and a manual steering task. A smooth transition from manual driving to assisted navigation is obtained thanks to a gradual visual servoing activation method that guarantees the continuity of the control law. The visual servoing task is then progressively activated, when the wheelchair gets closer to the walls or doorposts, in order to avoid collisions [43] . Experimental results clearly show the ability of the approach to provide an efficient solution for wall avoiding and doorway passing purposes [58] . This study is in the scope of the Inria large-scale initiative action PAL (see Section  8.2.6 ) as well as of the Apash project (see Section  8.1.1 ). Tests with disabled patients in the rehabilitation center Pôle Saint Hélier (Rennes) are under progress and first results prove the ability of our system to smoothly correct the trajectory of the wheelchair in case of hazardous situations.

Current research works are oriented towards multimodal sensor-based servoing, as well as haptic feedback that leads to an intuitive assistive wheelchair navigation. This work is realized in collaboration with Maud Marchal (Hybrid team). In addition, we are currently working with e-Motion team to design a vision-based human-aware semi-autonomous navigation system.

Social Spacing and human-robot interaction

Participants : Panagiotis Papadakis, Patrick Rives.

A novel probabilistic framework was introduced capable of instantiating diverse models of social spacing and accounting for distinctive dimensions in human-robot interaction, namely, perception capacity and certainty [42] . We have concretely shown how our method allows smooth adaptation in the situation awareness of a robot within common human-robot interaction examples and further showed its utility at the level of path planning by adapting trajectories to social sensitivity levels.

This approach is currently extended to take into account human activity cues in order to enhance robot mapping and navigation and in particular in filtering noisy human detections, detecting passages such as doors and staircases, inferring space occupancy and allowing navigation within unexplored areas.

Target tracking

Participants : Ivan Markovic, François Chaumette.

This study was realized in the scope of the FP7 Regpot Across project (see Section ) during the three-month visit of Ivan Markovic, Ph.D. student at the University of Zagreb. It consisted in developing a pedestrian visual tracking from an omni-directional fish-eye camera and a visual servoing control scheme so that a mobile robot is able to follow the pedestrian [37] . This study has been validated on our Pioneer robot (see Section  5.4.2 ).

Obstacle avoidance

Participants : Suman Raj Bista, Fabien Spindler, François Chaumette.

This study was realized in collaboration with Andrea Cherubini who is now Assistant Prof. at Université de Montpellier. It is concerned with our long term researches about visual navigation from a visual memory without any accurate 3D localization [4] . In order to deal with obstacle avoidance while preserving the visibility in the visual memory, we have proposed a control scheme based on tentacles for fusing the data provided by a pan-tilt camera and a laser range sensor [11] . A new study devoted to indoors navigation from segments has started recently.