Skip to main content

Motion Planning in Virtual Environments

Deliverables

A number of new algorithms have been developed to plan motions for entities with many degrees of freedom in complex scenes. These algorithms are very fast and flexible and can be used in many different domains, like virtual environments, robotics, and molecular biology.
New algorithms have been designed for planning manipulation tasks with articulated robots, possibly in combination with virtual characters. In particular techniques have been designed to deal with closed-loop systems. Such techniques can for example be used to plan the joint manipulation of objects by two or more robots.
A number of new algorithms have been developed for planning the motion of a group of entities in a virtual environment. Such entities must move as a group while avoiding obstacles. An important feature is that the group motion is natural. E.g. when moving through a narrow passage, the entities will wait for each other. The algorithms build a roadmap of possible motions, together with clearance information. In this way a corridor is created through which the group moves. Social potential fields are used to let the entities move as a group. Also techniques have been devised to add tactical constraints to the motion, e.g. to avoid dangerous places or to stay near to buildings.