January 25, 2022, 17:00 - 17:50
Rüdiger DillmannKIT/FZI Karlsruhe, Germany
Biomorphic Robot Controls: Event Driven Model Free Deep SNNs for Complex Visuomotor Tasks
The long term goal of this research within the framework of the European Human Brain Project (HBP) is to understand, to model and to translate biomorphic neural principles towards biocybernetic robot control systems. In comparison to conventional computing the brain is superior in terms of energy efficiency, robustness and adaptivity. Thus, we investigate into modeling biologic processes enabling the brain to perform sensomotoric computation and finally to implement it in silicon in form of biomorphic hardware. Todays neuromorphic hardware consists of spiking neural networks (SNNs) which can perform fast and efficient computations with continuous input - output streams based on synaptic plasticity. We focus on brain like senso-motor control tasks and ground them with the help of real robots. Spiking neural networks have the potential on replicating real neurons representing parts of their biological characteristics. SNNs are capable to perform synaptic spike-based communication with local brain functionalities supporting learning with the help of neural plasticity mechanisms. We assume, that the brain is forming sensor-motor primitives within building blocks composed for object detection, localization, event prediction, and finally the generation and execution of motion and interaction. The combination of neural motion primitives represent complex muscle motor synergies with the potential to learn complex large scale motions. Our SNN control architecture is capable to perform tasks like object recognition, object tracking, target reaching and grasping as well as collision- and obstacle avoidance. Closing the visuomotor loop by mapping the learned visual representation to motor commands show that SNNs learn without any planning algorithms nor inverse kinematics. SNNs are event driven and model free. We introduce deep continuous local learning mechanisms achieving state of the art robot accuracy on event stream benchmarks. Biologically plausible reward-learning rules based on synaptic sampling show that SNNs are capable of learning policies and various movement characteristics.
Links between reward-modulated synaptic plasticity and online reinforcement learning show proposing results. The hyper-parameters of this neuromodulation and their impact on performance are to be discussed with the help of some closed-loop sensorimotor experiments. The potential of deep reinforcement learning for target reaching affects object interaction, manipulation and grasping tasks and allows its realtime execution within dynamic situations. An event-driven binocular DVS system is used in stereo mode driven by micro saccades. The spiking feedback information from the DVS and from proprioception is mapped towards motion generating SNNs applying reward coupling and prediction error minimization techniques.
Future work towards the effective use of neuromorphic vision with emphasis to eye movement, micro saccades, visual affordance learning and high performance event prediction will be discussed. In addition it can be shown, that the brain-inspired computational paradigm can be extended towards SNN based navigation and mapping (BSLAM) forming episodic spatial neural memories with multi-scale learning capabilities. A software framework for developing and programming the related SNN-clusters and some complex biomorphic robot platforms are outlined.
Rüdiger Dillmann received his PhD from University of Karlsruhe in 1980. Since 1987 he has been Professor of the Department of Computer Science and Director of the Humanoids and Intelligence Systems Lab. at KIT. 2002 he became director of the innovation lab. IDS (interactive diagnosis systems) at the Research Center for Information Science (FZI), Karlsruhe. 2009 he founded the Institute of Anthropomatics and Robotics at the Karlsruhe Institute of Technology. His research interest is in the areas of human-robot interaction, neurorobotics with special emphasis on intelligent, autonomous and interactive robot behaviour generated with the help of machine learning methods and programming by demonstration (PbD). Other research interests include machine vision for mobile systems, man-machine cooperation, computer supported intervention in surgery and related simulation techniques. He is author/co-author of more than 1000 scientific publications, conference papers, several books and book contributions. He was Coordinator of the German Collaborative Research Center "Humanoid Robots" and several large scale European IPs. He is Editor in Chief of the book series COSMOS, Springer. Since 2018 he is Professor emeritus. He is now research director at FZI and is consulting start up companies and SMEs of his former PhD students.
He is IEEE- and IROS Fellow.
January 26, 2022, 11:00 - 11:50
Seul JungIntelligent Systems and Emotional Engineering(ISEE) Laboratory
Department of Mechatronics Engineering
Chungnam National University
Balancing Control in Systems
Maintaining balance in the nature is the most factor for human beings to keep our lives in peace and happiness. If the balance in nature breaks, human beings will be in danger and disaster. A human body itself is a balanced system. Likewise, balancing the system plays an important role in every dynamical machine and system. The balancing systems can be categorized into two parts: passive and active. The passive balance can be achieved from mechanical design such as center of gravity-based design and parallelogram. The active balance can be challengingly achieved by using various control algorithms. Each system requires an appropriate control algorithm to satisfy the balance from unstable condition. Therefore, the selection of appropriate balancing control algorithm is quite important in the system, which leads to the development of balancing control algorithms. In this talk, research from the inspiration of the balancing mechanism and control is introduced and shared. The balancing control of various systems is introduced in association with the appropriate control algorithms used in artificial intelligence, control and robotics areas.
Dr. Seul Jung was born in Incheon, Korea. He received the B.S. degree in Electrical & Computer Engineering from Wayne State University, MI, USA and the M.S & Ph.D. degrees from University of California, Davis, USA, both in Electrical & Computer Engineering.
Dr. Jung has joined the Department of Mechatronics Engineering at the Chungnam National University, Daejeon, Korea from 1997. While at Chungnam National University, he has engaged in research and teaching in the areas of intelligent systems, digital control, and signal processing. He has focused his efforts in robotics and intelligent control systems. He developed a graduate program in intelligent control and established an Intelligent Systems and Emotional Engineering (I.S.E.E) Laboratory for both theoretical and experimental studies. The Lab focuses on validation of theories by experimental studies of Mechatronics systems. He has published over 350 papers on these subjects. He is an author of 15 technical books about robotics, intelligent control, signal processing, and control system in Korean. His robotics research includes robot manipulator control, mobile robots, intelligent Mechatronics systems, and robot education. More recently, he has been engaged in developing control moment gyroscopes (CMGs) for Mechatronics systems. From 2002, he has organized the Creative and Intelligent Robot Contest (CIRO) for college student robotic competition and it continues until now.
Dr. Jung has been an active member in several IEEE Societies. He has served as a program committee member for many conferences. He also has been an executive office member in the Institute of Control, Robotics, and Systems (ICROS) in Korea. His important services include the organizing committee members and Program Chairs of many conferences. Now he serves as a general co-chair of ICCAS 2021.
January 26, 2022, 15:00 - 15:50
Mizuki OkaUniversity of Tsukuba, Japan
Exploring online social ecosystems through bio-inspired perspectives
Understanding how organisms reproduce and adapt to their environment and evolve is a significant issue in natural ecosystems. Similarly, online social systems such as social media are new ecosystems with complex functions. Recently, there has been a lot of attention on how much we can explain the ecosystem of social systems by using analogies of biological ecosystems because of the expectation that natural ecosystems provide a comprehensive theory for understanding social systems. In line with this research approach, I will discuss how the phenomena occurring in the online social ecosystem relate to living organisms' characteristics in this talk. By doing so, we hope to develop a strategy to transform online social systems into a constructive and democratic forum for making suggestions, debating, and communicating new ideas.
Mizuki Oka is an Associate Professor at the University of Tsukuba. She is also a head of the Artificial Life Research Group of the Japanese Society for Artificial Intelligence, an associate editor of Artificial Life Journal, a project manager of the IPA Exploratory IT Talent Search and Development Project, and a technical advisor to Blank Space Inc.
She conducts research on data analysis and utilization using machine learning, deep learning, and artificial life technologies. Based on her research at university, she focuses on the social implementation of new technologies and the provision of innovative value by bringing in unprecedented perspectives.