CCSS: FLASH: Drone Obstacle Avoidance with Event Cameras: Bio-Inspired Architecture, Algorithm, and Platform

CCSS：FLASH：使用事件相机进行无人机避障：仿生架构、算法和平台

• 批准号: 2302724
• 负责人: Longfei Shangguan
• 金额: $ 30万
• 依托单位: University of Pittsburgh
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2023
• 资助国家: 美国
• 起止时间: 2023-09-01 至 2026-08-31
• 项目状态: 未结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=2302724&HistoricalAwards=false
• 关键词: CCSS; FLASH; Drone; Obstacle; Avoidance

Drones hold immense potential as the future of deliveries, with the possibility of transporting food, medications, packages, and various other items directly to our doorsteps. However, realizing this vision requires overcoming various challenges, and obstacle avoidance stands as a critical component. A collision will not only cause financial loss but also threaten human safety. The objective of this research project is to develop an effective drone obstacle avoidance system by leveraging an innovative sensor called event camera. To achieve this goal, the PI systematically studied the mammalian visual system and proposes an innovative bio-inspired signal processing architecture, algorithm, and platform to fully unleash the potential of event cameras for drone obstacle avoidance. The education effort includes developing both individual and group-based research projects related to this proposal, actively recruiting female and underrepresented minority students for research, and providing research opportunities for K-12 students through outreach programs at Pitt. The overall research proceeds in three coherent thrusts. Thrust I: develop bio-inspired event filtering algorithms and architecture. Thrust II: develop bio-inspired event-matching algorithms. Thrust III: develop an open-sourced hardware-software interface. The developed solutions will lead to remarkable advancements in drone obstacle avoidance and bio-inspired designs. Moreover, we believe the proposed bio-inspired designs are generally applicable to other time-sensitive obstacle detection and localization applications such as autonomous driving. The developed open-sourced platform will allow mobile users and developers to fast prototype their own ideas without the need to build from scratch, which lowers the barrier of entry to drone-related research.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.

无人机作为未来的送货方式具有巨大的潜力，可以将食物、药品、包裹和其他各种物品直接送到我们的家门口。然而，实现这一愿景需要克服各种挑战，而避障是一个关键组成部分。碰撞不仅会造成经济损失，还会威胁到人类的安全。该研究项目的目标是通过利用一种名为事件摄像头的创新传感器来开发一种有效的无人机避障系统。为了实现这一目标，PI系统地研究了哺乳动物的视觉系统，并提出了一种创新的生物启发信号处理架构、算法和平台，以充分释放事件摄像头在无人机避障方面的潜力。教育工作包括开发与这一提案相关的个人和小组研究项目，积极招募女性和代表性不足的少数族裔学生进行研究，并通过PIT的外联计划为K-12学生提供研究机会。整个研究在三个连贯的推进中进行。主旨一：开发仿生事件过滤算法和体系结构。推力II：开发仿生事件匹配算法。推力三：开发一个开源的硬件-软件接口。开发的解决方案将导致无人机避障和仿生设计方面的显著进步。此外，我们相信拟议的仿生设计普遍适用于其他时间敏感的障碍物检测和定位应用，如自动驾驶。开发的开源平台将允许移动用户和开发人员快速构建他们自己的想法原型，而不需要从头开始构建，这降低了进入无人机相关研究的门槛。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。