C-BRIC e-Workshop
Event Date: | August 10, 2021 |
---|---|
Time: | 11:00 am (ET) / 8:00am (PT) AND 8:00 pm (ET) / 5:00 pm (PT) |
Priority: | No |
College Calendar: | Show |
Arijit Raychowdhury, Georgia Institute of Technology
Exploring Neural Dynamics for Designing the Next-Generation of Autonomous Systems
Exploring Neural Dynamics for Designing the Next-Generation of Autonomous Systems
Abstract:
As we enter an era of ubiquitous intelligence, powered by the recent advances in Deep Neural Networks and accelerated on custom silicon, research has started in earnest to explore the future of autonomous systems. It is enticing to ponder the role of neural dynamics and neurosciences to inform and drive the next generation of autonomy. In this talk, we will explore some of these questions with an emphasis on hardware and system design that mimics and emulates neural systems. As an example, we will look into NeuroSLAM – the first custom silicon demonstration of a coupled neuro-oscillator network, inspired by the brain of rodents, that can solve Simultaneous localization and mapping (SLAM) at breakthrough power efficiency and latency. In particular, low-power SLAM has gained increasing importance for its applications in power-limited edge devices such as compact unmanned aerial vehicles (UAVs), mixed-reality systems and small-sized vehicles. We will present a 7.25-to-8.79-TOPS/W mixed-signal neuromorphic SLAM accelerator in 65nm CMOS for applications in edge robotics. The oscillatory neural network emulates the brain’s spiking behavior and its continuous attractor property to achieve spatial cognition with a sparse energy distribution. Towards the end of the talk, we will explore end-to-end systems where the notion of event-driven, spike-based dynamics can be used as a unifying platform between cognition and control, providing end-to-end autonomy to hexapods. We will conclude the talk with some of our learnings from CBIRC and looking ahead at the future of autonomous systems.
Bio:
Arijit Raychowdhury is currently the Motorola Solutions Foundation Professor in the School of Electrical and Computer Engineering at the Georgia Institute of Technology where he joined in January, 2013. He received his Ph.D. degree in Electrical and Computer Engineering from Purdue University (2007) and his B.E. in Electrical and Telecommunication Engineering from Jadavpur University, India (2001). His industry experience includes six years as a Staff Scientist in the Circuits Research Lab, Intel Corporation, and a couple of years as an Analog Circuit Designer with Texas Instruments Inc. His research interests include low power digital and mixed-signal circuit design, design of power converters, sensors and exploring interactions of circuits with device technologies. Dr. Raychowdhury holds 27 U.S. and international patents and has published over 200 articles in journals and refereed conferences. He serves on the Technical Program Committees of ISSCC, VLSI Symp, CICC and DAC. He and his students have won many awards and fellowships including twelve best paper awards. He is currently serving as a Distinguished Lecturer of IEEE SSCS.