2018FLEX

A Holistic Approach Towards Flexible Hybrid Integration for Large-Area Sensor Platforms by System Design and Demonstrations (Room Spyglass)

15 Feb 18
8:45 AM - 9:05 AM

Tracks: 2018FLEX Full Conference, Emerging Technologies, IC/Hybrid Integration, Our Funded R&D Projects

Session 13: Emerging Capabilities

A Holistic Approach Towards Flexible Hybrid Integration for Large-Area Sensor Platforms by System Design and Demonstrations
Thursday, February 15, 2018
8:45 AM - 9:05 AM

Large-area flexible electronics is a natural platform for distributed sensing when the targets are on the scale of people and/or their infrastructure (meters and larger). This talk will focus on the critical issues of the architectural division of functions between the large-area and CMOS-IC domains, and approaches for the interfaces between the two domains. While the cost per unit area of thin film transistors is rapidly declining, TFTs have performance which is orders of magnitude worse than devices in modern CMOS IC’s. However, assembly cost, product flexibility and reliability will likely preclude placing an IC in every possible location electronics is desired. We will describe a holistic approach towards these trade-offs by examining a range of hybrid systems which have been designed and constructed at Princeton, including a flexible remote gesture-sensing and voice isolation sheet, a self-powered strain-sensing system for civil infrastructure, an EEG sensing and signal-processing cap, a handwriting-recognition sheet, a large-area sheet with wireless transceivers in wallpaper, and a pressure-sensing surface. In these systems the interfacing challenges have spanned many forms, including sensor-proximal amplification, harvester-localized power conversion, and signal modulation/demodulation for non-contact coupling. But, a guiding principle throughout has been minimizing the number of physical connections between the large-area and CMOS domains. This talk will look at how this can be done using specialized TFT circuits, such as low-power scan chains and LC oscillators, as well as algorithmically-driven TFT architectures, such as embedded compression and classification blocks, which exploit emerging algorithms from machine learning and statistical signal processing.