Human interaction with Motion Planning Algorithm
Résumé
This paper presents an interactive motion planning system to compute free collision motion in a numerical model. The system is based on interaction between a user and a motion planning algorithm. On one hand the user moves the object with an interactive device and on the other hand a motion planning algorithm searches a solution in the configuration space. The interaction aims at improving the guidance of an operator during a robot motion task in a virtual environment with the help of an automatic path planning algorithm. Existing works use a two-step decomposition which limits the interaction between the user and the ongoing process. We propose a modification of a classic motion planning method, the Rapidly-exploring Random Tree to build an Interactive-RRT. This method is based on exchanging pseudo-forces between the algorithm and the user, and on data gathering (labels) from the virtual scene. Examples are shown to illustrate the Interactive motion planning system with different interactive devices (space mouse and haptic arm). We analyze the influence of the user's dexterity to find a solution depending on various parameters of the algorithm and we show how we can adapt these parameters to a user.
Origine : Fichiers produits par l'(les) auteur(s)