This workshop has been published in IROS 2018.
Abstract
With state-of-the-art navigation algorithms, mobile robots can navigate safely among groups of people up to low-to-medium crowd density scenarios. When the density of people is not controlled, as in many daily situations (transport facilities, cultural or sports events, streets, shops, etc.), and when the level of density is high, most robots, when faced to human agents in their near vicinity, are programmed to stop to avoid collisions (the “freezing robot” problem). In those cases, the lack of static cues and the proximity of the human agents make it very challenging or even impossible to track and predict the motion of people around the robot, which in turns makes it hard for the robot to adapt its motion locally and to prevent contact.
However, a freezing robot becomes a static obstacle itself, further complicating the navigation problem for the humans around. For example, in evacuation scenarios, such an issue may elevate dangerously the evacuation times. Its navigation system should be designed in such a way that the robot is able to continue its task even in dense crowds. This objective goes well beyond state-of-the-art perception and control capabilities and probably requires completely new navigation, perception, cognition paradigms.
The central question of this workshop will be: How to overcome the perception and navigation problems mentioned above so that a mobile robot can engage in these crowded situations, and behave as safely as possible, so as to limit the number and types of contacts which may occur with the people in the space?
Details
- Title: From freezing to jostling robots: Current challenges and new paradigms for safe robot navigation in dense crowds
- Authors: Pettré, Julien; Babel, Marie; Hayet, Jean-Bernard; Salaris, Paolo; Salvini, Pericle
- Date of publication: 10/05/2018
- View article