Towards automatic generation of multimodal AR-training applications and workflow descriptions | IEEE Conference Publication | IEEE Xplore