Collaborative Research: CISE-MSI: DP: OAC: Integrated and Extensible Platform for Rethinking the Security of AI-assisted UAV Paradigm

无人机(UAV)被广泛采用来支持军事、民用和商业应用。随着人工智能技术的快速发展，人工智能辅助无人机范式在提高各种无人机操作的智能性、有效性和效率方面变得很有前途。与此同时，人工智能和无人机的融合也引发了新的安全担忧。该项目将调查和开发一个新的集成和可扩展的平台，以促进安全研究，发现漏洞并设计针对人工智能辅助无人机范例的对策。此外，该项目不仅将为加州州立大学圣马科斯分校--一所为少数族裔服务的机构--的研究人员建设和推广研究能力和资源，还将为大量未被充分代表的少数族裔学生提供研究和教育经验。该项目将首先设计和开发一个AI和无人机安全融合的研究平台。通过利用该平台，该项目将研究安全人工智能辅助无人机范例的两个方向，包括(1)启用人工智能的无人机异常检测的安全分析和(2)在对抗环境中稳健的人工智能辅助无人机操作。该项目的贡献将有三个方面：(1)深入了解人工智能和无人机集成带来的潜在漏洞，(2)探索构建有效攻击以利用漏洞并提供这些探索的影响的测量的可能性，以及(3)改进人工智能和无人机集成中流行的研究趋势的安全意识和防御方法。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。

Unmanned aerial vehicles (UAVs), or drones, are widely adopted to support a spectrum of military, civilian, and commercial applications. With the rapid development of AI techniques, the AI-assisted UAV paradigm becomes promising to enhance various UAV operations in terms of intelligence, effectiveness, and efficiency. At the same time, the integration of AI and UAVs also raises new security and safety concerns. This project will investigate and develop a novel integrated and extensible platform to promote security research that uncovers vulnerabilities and designs countermeasures toward the AI-assisted UAV paradigm. In addition, this project will not only build and promote research capacity and resources for researchers at California State University San Marcos - a minority-serving institution, but also provide research and education experience to a large number of underrepresented minority students. This project will first design and develop a research platform for the secure integration of AI and UAVs. By leveraging the platform, this project will then investigate two directions of the secure AI-assisted UAV paradigm, including (1) security analysis of AI-enabled UAV anomaly detection and (2) robust AI-assisted UAV operations in adversarial environments. The contributions of this project will be threefold: (1) obtain an in-depth understanding of potential vulnerabilities introduced by the integration of AI and UAV, (2) explore the possibility of constructing effective attacks to exploit vulnerabilities and provide the measurement of impacts of these explorations, and (3) improve the security awareness and defense approaches of the prevalent research trend in the integration of AI and UAVs.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.