W01 VLSI for IoT

Printer-friendly versionPDF version
Date: 
2017-03-31
Time: 
08:30-17:30
Location / Room: 
TBA

Organisers

Marilyn Wolf, Georgia Institute of Technology, US (Contact Marilyn Wolf)
Saibal Mukhopadhyay, Georgia Institute of Technology, US (Contact Saibal Mukhopadhyay)
Jörg Henkel, Karlsruhe Institute of Technology, DE (Contact Jörg Henkel)

This workshop focuses on the intersection between IoT and the VLSI systems required to build IoT systems. Internet-of-Things technology is being applied to a huge range of application areas. IoT systems make use of wireless networking to build physically distributed soft real-time systems. To fully deliver on the promise of IoT, we need to develop a new generation of VLSI devices for IoT nodes and hubs. IoT pushes VLSI system design in an entirely new direction - rather than concentrate on building the biggest, most complex chips, we need to develop a new generation of small, extremely inexpensive chips. This new generation of IoT devices will require advances in packaging, circuit design, low-power design, security, and system architectures.

The workshop features both invited speakers and submitted contributions from workshop participants. Interested participants can submit a 1-page summary of their proposed contribution by Friday, February 17, 2017 at https://easychair.org/conferences/?conf=vlsiiot17. Results will be announced by March 1, 2017.

Agenda

TimeLabelSession
08:30W01.1Devices and Systems

VLSI design - low power sensing and computation, wireless communication - defines the properties of IoT node devices. This session sets the stage for the day's discussion of VLSI for IoT. The first talk outlines how low power and low cost devices will influence the network architectures of IoT systems. The second talk proposes a fog-based paradigm for the IoT systems.

08:30W01.1.1Small Is Beautiful: The New VLSI Landscape
Marilyn Wolf, Georgia Tech, US

IoT requires small, capable devices. VLSI system design has traditionally pushed toward ever-bigger chips. IoT will require us to rethink our assumptions and develop new methods to design chips that are small, consume extremely low power levels, and provide the capabilities appropriate for the systems in which they opeate.

09:30W01.1.2Towards a Fog Computing Paradigm for Industrial Internet of Things
Paul Pop, Technical University of Denmark, DK

This talk proposes a distributed fog architecture for industrial IoT:

  • Requirements of applications in Manufacturing, Industrial automation, Robotics
  • Challenges in computation and communication
  • Fog Computing
  • Solutions: Virtualization (computation) and Time-Sensitive Networking, ultra-reliable 5G (communication)
  • (Resource management, provisioning)
10:30W01.2Architectures and Applications

This session looks at the architecture of IoT nodes and the characteristics of IoT applications. Hardware/software co-design techniques are well-suited to the design of low-power IoT nodes. Energy-oriented applications---smart grids and electric vehicles---offer key applications for IoT systems.

10:30W01.2.1Customizable Processor Acceleration of Time Series Similarity Search for the Internet of Things
Philip Brisk, University of California, Riverside, US

IoT nodes must both capture and analyze sensor data. This talk discusses the design of customizable processors for similarity search on sampled data.

11:15W01.2.2Deploying IoT for Energy Efficient Applications - Smart Grid and Electric Vehicles
Mohammad al Faruque, UC Irvine, US

IoT systems for applications that require energy efficiency.

14:15W01.3Circuits and Systems

Novel circuits enable new IoT applications. This session looks at two system designs for novel applications: personal health monitoring and neural-inspired image processing.

13:30W01.3.2Ultra-Low Power Neural Image Processing

Neuromorphic circuits for ultra-low power image processing.

14:15W01.3.1Toward Smart Wearable Systems for Personal Health Monitoring
Yu Cao, Arizona State University, US

Ultra-low power devices for personal health.

15:00W01.4SoC Design around ARM Cortex-M0 for Energy-efficient IoT Applications

Organiser:
Victor Nelson, Auburn University, US, Contact Victor Nelson


This talk explores various areas of IoT in the context of existing technologies and discusses features of the ARM Cortex-M family of CPUs suitable for applications interesting to IoT. The talk will be followed by a demo on prototyping an IoT application built on the framework of the SoC Design Education Kit from ARM University Program using the energy-efficient ARM Cortex-M0 DesignStart IP core.


15:00W01.4.1SoC Design around ARM Cortex-M0 for Energy-efficient IoT Applications: Techniqwues

This talk describes the range of features of the ARM ecosystem that enable IoT.

15:30W01.4.2SoC Design around ARM Cortex-M0 for Energy-efficient IoT Applications: Demonstrations
Victor Nelson, Auburn University, US

A demonstration of the capabilities of ARM systems for IoT applications.

16:45W01.5Participant Presentations and Discussions

Presentations from workshop participants and group discussion of workshop topics.