Neuron Inspired Architectures
Main Lab Location:
CiNet (Main bldg.)
・Guest Professor, Osaka University, Graduate School of Information Science & Technology
・Visiting Professor, University of Hyogo, Graduate School of Engineering
・Visiting Professor, Kobe University, Graduate School of Health Sciences
・Member of International Technology Roadmap for Semiconductors (ITRS), Emerging Research Devices (ERD) working group
1-4 Yamadaoka, Suita City Osaka, 565-0871
I do research on communication and computation systems composed of extremely simple elements that derive their functionality from interactions with each other. Such highly distributed systems are easy to manage, low in energy consumption, robust to defects of elements, and cheap to manufacture, yet they will allow users to gain a powerful control and understanding of their environment. Neural systems have been my special research interest, because of their innate ability to represent and process information in real time through simple interactions.
My lab aims to develop wireless sensor networks that consist of huge numbers of extremely simple nodes, each being at most 1 mm in any dimension, consuming only power provided to it by its environment, and costing less than 1 Yen to mass manufacture. Key requirements are the use of pulse-based signaling both inside nodes as well as between nodes. Nodes collaborate to obtain information about the environment by combining pulses from other nodes through strategies resembling those found in the brain, like synchronization
desynchronization of pulse trains and the use of spatial temporal activation patterns.
The pulse-based wireless sensor networks we aim to develop will be employed cost-effectively in much higher densities than currently possible by conventional wireless sensor networks. This should allow much faster and precise detection of anomalies in the environment (such as chemical disasters), as well as real-time monitoring. Due to the small size and low cost of nodes, no retrieval of nodes after use is necessary, and their energy harvesting ability will allow them to be active for extended periods of time without maintenance. New applications become possible, like attaching sensor nodes to insects, spreading sensor nodes in the air, and applying sensor nodes to surfaces.
I am also active in research on Cellular Automata-based Nanocomputers as next generation computer architectures. Like pulse-based wireless sensor networks, Cellular Automata are highly distributed systems that employ collaboration between simple elements to obtain a powerful functionality. The potential of Cellular Automata lies in their regular structure, which allows manufacturing through directed self-assembly combined with advanced lithographic techniques.
Peper, F., Wakamiya, N., Kasamatsu, A., Tanaka, S., Leibnitz, K., Teramae, J., Kasai, K., Otomo, A. On Neuro-inpired Wireless Sensor Networks. Proc. 10th Int. Conf. on Ubiquitous Intelligence & Computing (UIC 2013), Vietri sul Mare, Italy, 589-594 (Dec. 2013)
Peper, F., Lee, J., Carmona, J., Cortadella, J., Morita, K. Brownian Circuits: Fundamentals. ACM Journal on Emerging Technologies in Computing Systems 9 (1), 3:1-24 (2013)
Bandyopadhyay, A., Pati, R., Sahu, S., Peper, F., Fujita, D. Massively parallel computing on an organic molecular layer. Nature Physics 6 (5), 369-375 (2010)
Peper, F., Lee, J., Abo, F., Isokawa, T., Adachi, S., Matsui, N., Mashiko, S. Fault-tolerance in nanocomputers: a cellular array approach. IEEE Trans. Nanotechnology 3 (1), 187-201 (2004)
1989 Ph.D. in Theoretical Computer Science at Delft University of Technology, the Netherlands