II-NEW: Collaborative Research: An Extensible Software Infrastructure for Unmanned Aerial Vehicles

无人机是一种新兴的计算平台，在第三世界国家具有商品递送、交通控制、新闻报道、自然灾害管理、移动社交网络和互联网连接等多种应用。展望未来，无人机令人兴奋的潜力伴随着重大障碍，需要广泛和协调的跨学科研究，重点放在实时系统设计、能效、安全和安保、可编程性、机器人学和机械设计等方面。不幸的是，今天的大多数无人机系统要么是专有的，要么是专门针对航空/机器人任务的特定目标的。在这个项目中，我们将探索为无人机开发一个开源和可扩展的软件基础设施的可行性和社区支持，以促进对这项前景光明的技术的研究和教育，最终导致更好地理解、评估，并可能更广泛地应用无人机。PI将与学术界和工业界的领先专家合作组建一个社区。这一早期探索将导致第一个基础设施的开发计划，以允许跨计算堆栈、跨操作系统(OS)、跨虚拟机(VM)、编译器、编程语言和应用程序的可扩展无人机软件设计。如果成功，计划中的基础设施可能会显著转移无人机系统最先进的重点，在无人机系统中，精细化的支持仅限于硬件驱动程序和机器人控制，OS/VM支持是原始的，应用程序编程接口(API)最少。此外，该基础设施将为无人机教育提供一个平台，使学生能够在大学水平的课程中更早地学习无人机。

Unmanned aerial vehicles (UAVs) are an emerging computing platform becoming increasingly common in our society, with diverse applications such as merchandise delivery, traffic control, news reporting, natural disaster management, mobile social networks, and Internet connectivity in third-world countries. Looking forward, the exciting potential of UAVs is accompanied with significant hurdles that call for broad and concerted interdisciplinary research, with diverse foci on real-time system design, energy efficiency, safety and security, programmability, robotics and mechanical design, among others. Unfortunately, the majority of UAV systems today are either proprietary, or specific to particular goals of aviation/robotic missions. In this project, we will explore the feasibility and community support of developing an open-source and extensible software infrastructure for UAVs to promote research and education of this promising technology, ultimately leading to a better understanding, assessment, and possibly wider application of UAVs. The PIs will work with leading experts in both academia and industry in assembling a community.This early-stage exploration will result in a development plan for a first infrastructure to allow for extensible UAV software design across the computing stack, spanning operating systems (OS), virtual machines (VM), compilers, programming languages, and applications. If successful, the planned infrastructure may significantly shift the focus of the state of the art of UAV systems, where refined support is limited to hardware drivers and robotics control, OS/VM support is primitive, and Application Programming Interface (API) is minimal. Additionally, the infrastructure will provide a platform for UAV education, which will allow for students to study UAVs at a significantly earlier stage in a college level curriculum.