Day 1 Speakers: Sensors for the Internet of Things

Prof. David Blaauw

Kensall D. Wise Collegiate Professor
University of Michigan

"Sensors for the IoT"

The internet of things (IoT) at the edge of the cloud is rapidly evolving, and is poised to become a large market for the semiconductor industry. Most IoT devices do not require man-machine interfaces, allowing their sizes to be reduced to the mm-scale.  However, this means that their primary input data comes from sensors.  High-performance sensors and sensor interfaces are thus of utmost importance to their successful implementation. However, IoT devices face unique challenges:  their power supply is often poorly regulated, their operating temperatures can change rapidly and range widely and their small size limits the sensor technologies that can be employed.  In this talk, we discuss how to overcome these challenges and present a number of sensor interfacing solutions that enable small (down to mm-scale) IoT devices to perform accurate sensing of their environment. In particular, we will discuss pressure, temperature, light and pH sensing, as well as more complex sensing modalities such as audio and image acquisition.  Finally, we will showcase a number of complete mm-scale sensing systems operating in deployed environments.

Dr. Patrick P. Mercier

Associate Professor
Center for Wearable Sensors at UC San Diego

"Energy-Efficient Power Management Circuits for IoT and Wearable Devices"

Next-generation wearable and IoT devices require small form factor implementations with long battery life. Most of their circuitry, including computation, sensing, and wireless communications, will be realized in SoCs that will operate at 1V or below. However, small batteries typically provide higher voltages than this, and so dc–dc converters are required.  These must then be small, and capable of efficient operation over a wide range of loads (nanowatts for always-on circuits to milliwatts for duty-cycled radios). Unfortunately, dc-dc converters often require bulky inductors, and scaled CMOS processes cannot handle the required voltages. This presentation will discuss topological solutions and circuit optimizations to increase the efficiency of Li-ion-to-SoC dc-dc converters, while minimizing their quiescent power and reducing their size. Energy harvesting solutions, including techniques for low-area multi-modal harvesting power aggregation, will also be discussed.  

Prof. Farrokh Ayazi

Ken Byers Professor in Microsystems
Georgia Institute of Technology

"Low-Power Inertial Measurement Chips for Health Informatics and IoT"

This talk describes the development of contact microphones with micro-g resolution and multi-axis gyroscopes with self-calibration capabilities for use as acoustic auscultation devices in body-worn sensor arrays.  Combining such devices into a multi-degree-of-freedom inertial measurement unit (IMU) on a single-chip, enables the simultaneous measurement of cardiopulmonary sounds, chest wall motion, heart ballistocardiogram signals, as well as of user body motion.  The CMOS ASIC consists of switched capacitor and transimpedance amplifier front-end circuits that utilize correlated double sampling and chopping for the dynamic cancellation of offset and flicker noise, and use charge injection calibration techniques to compensate for MEMS capacitor mismatch. We discuss the prospects of reducing power while maintaining precision in interface ICs for MEMS IMUs.

Mr. Luca Sant

Infineon Technologies Austria AG

"Monolithic capacitive pressure sensors in standard CMOS technology"

The increasing demand for environmental sensors in mobile applications such as phones and smart-watches has driven the development of low-power barometers, which must achieve centimeter-level accuracy in tiny packages. Building on existing MEMS microphones, this paper describes the development of a high-performance capacitive pressure sensor. Striving for minimum area and cost, the MEMS sensor array was designed for co-integration with the read-out circuitry, and was fabricated in a standard 0.13μm CMOS technology. To meet the requirements of different applications, its accuracy can be traded off with power consumption. In high precision mode, better than 0.5PaRMS accuracy (equivalent to less than 5cm) can be achieved at a refresh rate of 1Hz, while drawing approximately 32μA. This matches the performance of state-of-the-art barometric pressure sensors based on piezoelectric sensors.  

Dr. Nick Van Helleputte

R&D manager of the Connected Health Solutions group
Analog R&D Design Engineer

"Wearable and Non-Contact Health Sensing"

For a number of chronic cardio-pulmonary diseases, remote home monitoring via wearables is rapidly gaining popularity. This talk will focus on recent developments in this field. Multiple sensor modalities targeting ever more complex vital-sign recording are being embedded into single devices. At the same time, ever more functionality is being embedded into miniature ASICs enabling extremely small devices that can operate at very low power. Besides wearables, the talk will also discuss new sensing technologies that enable vital-sign sensing without requiring any physical contact to the human body. Such technology can be embedded in our surroundings to enable truly unobtrusive continuous health monitoring.