Modeling of motion primitive architectures using domain-specific languages
Arne Nordmann · PUB – Publications at Bielefeld University (Bielefeld University) · 2016
Many aspects in robotics, and their omnipresent ideal models, animals and humans, are still not understood or explored well enough, for example producing motions of animal-and human-like complexity.To explore the inner workings of systems studying this complexity, the essential concepts of interest need to be made explicit and raised from the code-level to a higher level of abstraction to be able to reason about them.This work introduces a model-driven engineering approach for complex movement control architectures based on motion primitives, which in recent years have been a central development towards adaptive and flexible control of complex and compliant robots.The goal is to realize rich motor skills through the composition of motion primitives.This thesis proposes a design process to analyze the control architectures of representative example systems to identify common abstractions.Identified and formalized concepts can then be used to automate software development of motion primitive architectures through model-driven engineering methods and domain-specific languages.It turns out that the introduced notion of motion primitives implemented as dynamical systems with machine learning capabilities, provide the computational building block for a large class of such control architectures.Building on the identified concepts, a set of modularized domain-specific languages allows the compact specification of motion primitive architectures.This paves the way for domain experts rather than computing technology specialists to produce systems, which is one of the main goals of this work.The approach and the accompanying model-driven engineering toolchain is evaluated in a task of the European Robotics Challenges (EuRoC) and a real world example of automatic laundry grasping with the KUKA Lightweight Robot IV, where executable source-code is automatically generated from the domain-specific language specification.