Embedded AI Systems for Autonomous Driving and Smart Health

Abstract

Embedded Artificial Intelligence is rapidly emerging as a transformative computing paradigm, enabling intelligent, real-time, and privacy-preserving interactions with the physical world. In this talk, I will present our recent work on Embedded AI, including real-time inference on resource-constrained platforms, and AI-empowered systems for autonomous driving and smart health.

First, I will introduce a novel real-time deep learning framework that integrates model architecture optimization with real-time scheduling to enable the concurrent execution of multiple deep learning tasks. I will then discuss Soar, the first end-to-end intelligent roadside infrastructure system designed to enhance the safety of autonomous driving. Deployed on the campus of CUHK, this system is the first real-world open testbed that showcases the potential of infrastructure-assisted autonomous driving. Next, I will discuss several new systems that achieve real-time 3D collaborative perception and high-definition (HD) mapping for autonomous driving, with centimeter-level accuracy. Finally, I will introduce our work on integrating multi-modal sensors, federated learning algorithms, and foundation models to monitor multidimensional digital biomarkers associated with aging-related degenerative diseases. This system addresses key challenges such as limited data labels, data heterogeneity, and constrained computational resources. In collaboration with medical experts and local hospitals, we are currently deploying this system in large-scale clinical trials involving over 1,000 subjects.

About this Lecture

Request this Lecture

To request this particular lecture, please complete this online form.

Request a Tour

To request a tour with this speaker, please complete this online form.

All requests will be sent to ACM headquarters for review.