Dynamic motion learning for multi-DOF flexible-joint robots using active–passive motor babbling through deep learning

This paper proposes a learning strategy for robots with flexible joints having multi-degrees of freedom in order to achieve dynamic motion tasks. In spite of there being several potential benefits of flexible-joint robots such as exploitation of intrinsic dynamics and passive adaptation to environmental changes with mechanical compliance, controlling such robots is challenging because of increased complexity of their dynamics. To achieve dynamic movements, we introduce a two-phase learning framework of the body dynamics of the robot using a recurrent neural network motivated by a deep learning strategy. The proposed methodology comprises a pre-training phase with motor babbling and a fine-tuning phase with additional learning of the target tasks. In the pre-training phase, we consider active and passive exploratory motions for efficient acquisition of body dynamics. In the fine-tuning phase, the learned body dynamics are adjusted for specific tasks. We demonstrate the effectiveness of the proposed methodology in achieving dynamic tasks involving constrained movement requiring interactions with the environment on a simulated robot model and an actual PR2 robot both of which have a compliantly actuated seven degree-of-freedom arm. The results illustrate a reduction in the required number of training iterations for task learning and generalization capabilities for untrained situations.     

Timing Analysis of Real-Time Communication Under Electromagnetic Interference

This paper discusses aspects of dependability of real-time communication. In particular, we consider timing behaviour under fault conditions for Controller Area Network (CAN) and the extension Time-triggered CAN (TTCAN) based on a time-driven schedule. We discuss the differences between these buses and their behaviour under electromagnetic interference. We present response timing analyses for CAN and TTCAN in the presence of transient network faults using a probabilistic fault model where random faults from electromagnetic interference occur. The CAN analysis provides a probability distribution of worst case response times for message frames. The results indicate that CAN may generally provide a higher probability of delivering messages on time than TTCAN. The CAN analysis result is used to discuss an approach to implementing a bus guardian for event-triggered systems.Ian Broster is a research associate at the University of York, his research includes real-time communication and work on the CAN protocol. Current research focuses on next-generation flexible scheduling for real-time operating systems. His research interests include probabilistic analysis, timing analysis of non-deterministic systems, flexible scheduling, real-time communication, simulation and modelling. He received his M.Eng. degree in 1999 and a Ph.D. in 2003 for his work on flexible real-time communication at the University of York, U.K.Alan Burns has worked for many years on a number of different aspects of real-time systems engineering. He graduated in 1974 in Mathematics from Sheffield University; he then took a D.Phil, in the Computer Science Department at the University of York. After a short period of employment at UKAEA Research Centre, Harwell, he was appointed to a lectureship at Bradford University in 1979. He was subsequently promoted to Senior Lecturer in 1986. In January 1990 he took up a Readership at the University of York in the Computer Science Department. During 1994 he was promoted to a Personal Chair. In 1999 he became Head of the Computer Science Department at York.Guillermo Rodríguez-Navas holds a degree in Telecommunication Engineering by the University of Vigo, Spain. He is currently doing a Ph.D. in Computer Science at the University of the Balearic Islands, Spain. He is also a member of the System, Robotics and Vision (SRV) research group at this university. His research is focused on dependable and real-time distributed embedded systems. In particular, he has addressed various issues related to the Controller Area Network (CAN) field bus, such as fault tolerance, clock synchronization and response time analysis.