Omnidirectional photometric visual path following for wheelchair autonomous driving

Youssef Alj (1) , Guillaume Caron (1) , Nicolas Ragot (2, 3)

1 MIS - Modélisation, Information et Systèmes - UR UPJV 4290
2 Pôle Instrumentation, Informatique et Systèmes
3 IRSEEM - Institut de Recherche en Systèmes Electroniques Embarqués

In this paper we address the issue of autonomous wheelchair navigation. Using an omnidirectional camera, we propose a system that allows wheelchair driving using a visual path following technique. First, a visual path composed of target images is acquired. Second, successive visual servoings are performed by minimizing the error between the current and a target image. Experiments were conducted on a variety of paths: straight line, curved line and forward/backward movements. The ground truth is obtained using a Vicon system that records wheelchair trajectories during the learning and the path following steps.

Domaines: Vision par ordinateur et reconnaissance de formes [cs.CV]; Robotique [cs.RO]

Youssef Alj, Guillaume Caron, Nicolas Ragot. Omnidirectional photometric visual path following for wheelchair autonomous driving. 1st Healthcare Technology Days, CARETECH, Dec 2014, Rouen, France. ⟨hal-01281652⟩