A new model to design software architectures for mobile service robots

Both Robots and Personal Computers established new markets and became mass-products in the mid-1970s. They were enabling factors in Automation and Information Technology respectively. However, while you can see Personal Computers in almost every home nowadays, the domain of Robots is mostly restricted to industrial automation. Due to the physical impact of robots, a safe design is essential which most manipulators still lack of today and therefore prevent their application for personal use. A slow transition can be noticed however by the introduction of dedicated robots for specific tasks in environments shared with humans. These are classified as service robots. TUM’s Department for Robotics and Embedded Systems approach to service robotics was driven by several real world application scenarios for autonomous mobile robots in life science laboratories, changeable factories, TV studios and educational as well as domestic application domains. Opposed to manipulators for industrial automation, most service robots carry much more sensor equipment and computing power to perceive their environment and to process the acquired sensor data for autonomous behaviour. The variety of utilised hardware and the versatile use cases for service robots turn them into complex, heterogeneous, and distributed IT systems. To avoid inventing custom software architectures for every newly created service robot, standardisation of software components and interfaces is key for their development. This thesis proposes a novel model to classify the hardand software components of autonomous service robots and discusses their interfaces, generalisations, and specialisations. A large part of this work is dedicated to the design and implementation of perception modules as they are essential for service robots. In summary, the model covers the sensors, the actuators and the corresponding busses and networks as well as the overlying software counterparts for the communication channels, device classes, and the software components for perception, task planning, and applications. The result section discusses its successful application in state of the art projects developed at our department.