Intuitive programming and robot skills
There is an ongoing shift in industry from mass production to low-batch production with highly individualized goods.
This increases the effort to program robots, which is typically carried out by robot experts. For keeping the production economical, new programming approaches are required that allow non-expert human coworkers to instruct robots easily.
One approach to make programming more intuitive is to use robotic skills that are preprogrammed, however flexible, software modules that only need to be parametrized. Robot skills contain all information, behavioral logic, recover strategies, controllers, etc., which are needed to execute a certain task such as drilling or screwing. The non-expert can sequence single skills to obtain a complex autonomous or interactive robot task while only setting a limited number of parameters.
At DLR, a skill software architecture has been developed which is robust, fast to parameterize, and can handle dynamic, physical human-robot interaction. For multimodal interaction, intuitive user interfaces have been developed.
In robotics, a paradigm shift took place towards designing robots capable of directly interacting and collaborating with humans closely in both industrial and domestic environments.
Possible applications range from assembly assistance in car production to providing help to disabled and infirm people in their households. In this endeavor, various challenges are to be tackled, which is why human-robot interaction is a vivid field of research.
Over the years, DLR has gained much experience in the development and control of human-friendly robots, and the design of applications that allow safe coexistence. At Automatica, novel methods for intuitive programming, hardware designs for an optimized robot workspace as well as a mechanical safety mechanism are presented.