Special Issue on Humanoid Robot Applications in Real World Scenarios, IEEE Robotics and Automation Magazine.
Humanoids fascinate through their anthropomorphic appearance, high dexterity and the potential to work in human-tailored environment and to interact with humans in novel applications. In our research, we promote two real-world applications in physiotherapeutic juggling and assisted walking using two compliant humanoids COMAN and COMAN+. We focus on rehabilitation, which as a result of changing demographics becomes an increasingly crucial application field. But alike most humanoid experiments, the realization of these scenarios is challenging due to the fact that the hardware is brittle, software is complex and control remains highly demanding. In this article, we describe an integrative and transparent control architecture that alleviates this complexity by strictly adhering in design and implementation to a component-based approach. It promotes flexibility and reusability and allows transparent switching between different humanoid robots, between simulation and real world, between control paradigms. It also orchestrates the integration of real-time and non-real-time components, including a Virtual Reality framework, towards rich user interaction.
Pouya Mohammadi 
Enrico Mingo Hoffman 
Niels Dehio 
Milad S. Malekzadeh 
Martin Giese 
Nikos G. Tsagarakis 
Jochen J. Steil 
 The Institute for Robotics and Process Control (IRP), Technische Universität Braunschweig, Germany
 Advanced Robotics Department (ADVR), Istituto Italiano di Tecnologia, Genova, Italy
 Section for Computational Sensomotorics, Department of Cognitive Neurology, HIH, CIN, University Clinic Tübingen, Germany
 Intelligente Prozessautomation und Robotik (IPR), Karlsruher Institut für Technologie, Germany