Providing a drilling robot with the right instructions

This paper presents the concept and implementation of a task-level instruction system for an autonomous drilling robot used in railway installation work. The research objectives are: to reduce the required programming effort, make more intelligent robot behaviour possible and bring the programming abstraction level to a task-instruction level. Relevant railway construction knowledge and task related information is stored in the drilling robot using product models. Standardised exchange formats for product models enables re-use of CAD information by the drilling robot which significantly reduces the required robot instruction effort.