Architecture of a cyberphysical avatar

• Published in: Proceedings of the ACM/IEEE 4th International Conference on Cyber-Physical Systems pp. 189 - 198
• Main Authors: Han, Song, Mok, Aloysius K., Meng, Jianyong, Wei, Yi-Hung, Huang, Pei-Chi, Leng, Quan, Zhu, Xiuming, Sentis, Luis, Kim, Kwan Suk, Miikkulainen, Risto
• Format: Conference Proceeding
• Language: English
• Published: New York, NY, USA ACM 08.04.2013; IEEE
• Series: ACM Conferences
• Subjects: Avatars; Computer systems organization > Embedded and cyber-physical systems > Robotics; Computing methodologies > Artificial intelligence > Control methods > Robotic planning; Computing methodologies > Artificial intelligence > Planning and scheduling > Robotic planning; Computing methodologies > Machine learning > Machine learning approaches > Neural networks; Grasping; Networks > Network architectures; Neural networks; Real-time systems; Robot kinematics; Robot sensing systems
• ISBN: 1450319963; 9781450319966
• DOI: 10.1145/2502524.2502550

This paper introduces the concept of a cyberphysical avatar which is defined to be a semi-autonomous robotic system that adjusts to an unstructured environment and performs physical tasks subject to critical timing constraints while under human supervision. Cyberphysical avatar integrates the recent advance in three technologies: body-compliant control in robotics, neuroevolution in machine learning and QoS guarantees in real-time communication. Body-compliant control is essential for operator safety since cyberphysical avatars perform cooperative tasks in close proximity to humans. Neuroevolution technique is essential for "programming" cyberphysical avatars inasmuch as they are to be used by non-experts for a large array of tasks, some unforeseen, in an unstructured environment. QoS-guaranteed real-time communication is essential to provide predictable, bounded-time response in human-avatar interaction. By integrating these technologies, we have built a prototype cyberphysical avatar testbed.