Reactive assembly planning in a dynamic virtual environment

Working in the virtual environment has the advantages of unsurpassed support from computer tools. However, its success in industry applications still relies on how accurately a virtual environment represents the real life activities. Presented in this paper is a method of reactive assembly planning that is able to handle environmental uncertainty. It uses automatic object modeling to update the virtual environment according to the change in a working environment, and a biologically inspired neural network to reactively generate collision-free paths at real time.