Abstract: Weight is an impediment to drone flight. Greater intelligence implies more powerful computing and richer sensing. This implies heavier CPUs, GPUs, memory, sensors and batteries. Unlike an autonomous road vehicle that easily carries LiDAR and multiple video cameras, the lightest drones only carry a single video camera on an actuable gimbal as well as minimal on-board compute. Within these severe constraints, can we achieve state-of-the-art (SOTA) AI on commercial off-the-shelf (COTS) ultralight drones? Specifically, can we achieve fully autonomous drone flight without a human pilot? If feasible, autonomy would no longer be an exotic feature only available on large, heavy and expensive high-end drones, but a basic attribute of virtually all drones. The use of COTS drones would drive down cost, increase availability, and generate new business models by eliminating the need for a skilled and expensive human pilot. This vision is being explored in the SteelEagle project at Carnegie Mellon University. Our approach uses a hardware component called a DroneHub to bridge the unmodified on-board drone hardware and software with the more powerful computing resources available via 5G on a ground-based cloudlet. The DroneHub also supports transient disconnected operation in case of brief disruptions of the wireless network. Our experimental results show that this vision is indeed achievable, but its real-world success will require better DroneHubs than are available today. We explore these concepts in this talk, with the goal of identifying the attributes of an ideal DroneHub for the future.

Bio: Satya's multi-decade research career has focused on the challenges of performance, scalability, availability and trust in information systems that reach from the cloud to the mobile edge of the Internet. In the course of this work, he has pioneered many advances in distributed systems, mobile computing, pervasive computing, and the Internet of Things (IoT). Most recently, he has been viewed as "The Father of Edge Computing" for his seminal 2009 paper, and his pioneering contributions to the foundations of edge computing. Satya is the Jaime Carbonell University Professor of Computer Science at Carnegie Mellon University. He received the PhD in Computer Science from Carnegie Mellon, after Bachelor's and Master's degrees from the Indian Institute of Technology, Madras. He is a Fellow of the ACM, IEEE, and EAI, and a member of the National Academy of Engineering.

Speakers:
Xiaomin Ouyang (HKUST)
Akshay Gadre (University of Washington)
Dong Ma (University of Cambridge)
Jun Han (KAIST)
Luca Mottola (Politecnico di Milano)

Abstract: In an era of rapid technological evolution, accelerated by advances in AI and machine learning, endpoints such as mobile devices and IoT devices such as smart sensors have become powerful tools for easily manipulating digital content, be it images, video, speech, or other sensor data. As a result, it is increasingly difficult to determine whether the content consumed today by humans and machines is genuine or fake, contributing to the rise of digital abuse and misinformation in society. Content provenance technology, promoted by the Coalition for Content Provenance and Authenticity (C2PA), aims to address these challenges. This talk will provide an overview of standards work and ecosystem adoption of content provenance and highlight how trust can be further strengthened through secure IoT platforms that attach provenance information right at the source of capture.

Short Bio: Yogesh Deshpande is a Principal Security Architect in Arm, Cambridge, UK. He works on developing various Security Standards and Open-Source Implementation, with primary focus on Remote Attestation, Content Provenance and Supply Chain Security.

This session highlights the 25th anniversary of IEEE Pervasive Computing and celebrates the community, ideas, and systems work that shaped the field.