NSF-AoF: FET: Small: Ubiquitous in-sensor computing for adaptive intelligent systems

Miniaturized sensor systems with built-in memory and computing functionalities are the cornerstones of artificial intelligence at the edge. However, in currently deployed systems, sensing and computing occur in separate physical locations, imposing massive data shuttling between the sensor module and the cloud-computing platform. This is unsustainable in terms of energy efficiency, latency, and capacity to process sensor data, and hence has a negative environmental impact with billions of sensors connected in the era of the Internet of Things (IoT).

The proposed project intends to go beyond state-of-the-art by system-level integration of sensing, memory, and computing functionalities into one chip, allowing for ubiquitous applications at low energy budget and low latency. Furthermore, building such systems on flexible substrates will enable affordable and biodegradable smart-wearables electronics capable of monitoring human health continuously and adaptively.

The proposed educational and outreach activities will promote STEM careers, encourage diversity in engineering education and research, and significantly impact securing the future prosperity of the U.S. and the European collaborative partner (Finland).

The proposed research aims at delivering intelligent and energy-efficient wearable electronics that will become ubiquitous in the era of IoT. The specific objectives towards this goal are as follows: 1) to design and fabricate emerging materials and devices for flexible sensors; 2) to integrate ferroelectric sensors and memristor crossbar arrays into a flexible near-sensor computing system with embedded security functionality; and 3) to demonstrate an in-sensor computing platform where emerging devices will be used as both sensing and non-volatile memory elements for in-pixel intelligent processing of images.

The proposed research will enable the next-generation smart and flexible wearable electronics to process the acquired information onsite. By integrating the sensor module with the computing engine, the emerging hardware technologies will substantially improve power efficiency and computing throughput.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.