Enabling and Securing Robotic Team Situational Awareness
Abstract
It is envisioned that the principal Army unit operating in 2050 will be mixed human-robot teams. Companion robot will be pervasively deployed to collaborate with human members in mixed teams to provide intelligence, surveillance, and reconnaissance without putting soldiers in dangerous situations. Robotic Team Situational Awareness (TSA) addresses the problem of how robots in a mixed human-robot team can perceive human teammates, understand individual behaviors that reflect team roles, and interpret team intents that reflect mission objectives. As an enabling capability for robot decision making, cognition, and interaction with human teammates, robotic TSA is essential to successful accomplishment of ongoing missions. Enabling robotic TSA and securing robotic TSA are two essential and strongly integrated computational capabilities that must be ensured in order for companion robots to effectively collaborate with soldier teammates, especially in an inherently adversarial and highly dynamic environment, such as tactical battlefields. Robotic TSA must be automated and resilient in highly dynamic tactical battlefields, where soldiers are generally not available to elaborate their behaviors or intents to robots when faced with enemy attacks or during an emergency. Meanwhile, robotic TSA must be protected against misinformation attacks, which are the major security threat to robotic TSA, where an adversary intends to deceive a robot using false or inaccurate information. Misinformation as a weapon is especially powerful and stealthy in attacking mixed human-robot teams, which can render fundamental robotic TSA functionalities inaccurate and unreliable, and even cause devastating damages to a team such as claiming the lives of soldiers and ruining critical missions. The goal of this research project is to investigate algorithms and approaches that enable and secure robotic TSA for mixed human-robot teaming in data-rich, dynamic, and adversarial environments. Specifically, we propose to achieve three research objectives: (1) Multisensory data fusion for robotic TSA, which aims to develop robotic TSA methods that can integrate multisensory data for robots to interpret individual behaviors and team intents; (2) Robustness quantification of robotic TSA under misinformation attacks, which aims to quantify the robustness of robotic TSA methods by systematically investigating their resistance to misinformation attacks; and (3) Active security design of robotic TSA, which aims to integrate security protection into the design of robotic TSA and decision making methods to make them robust against misinformation attacks. Our proposed project is innovative and will advance the state-of-the-art research on mixed human-robot teaming. We are the first to investigate the essential computational problem of robotic TSA from the fresh perspective of actively integrating security protection to enable secure robotic TSA, thus accelerating the significant trend towards mixed human-robot teaming in adversarial tactical environments. The proposed algorithms of multisensory information integration to enable robotic awareness of human behaviors and team intents are novel. Our proposed research on enhancing security protection against misinformation attacks in both physical and cyber spaces with high-dimensional multisensory data is innovative. More importantly, this project will open up a new cross-domain research area between human-robot teaming and cyber security, and provide a principled understanding of secure robotic TSA, which has not yet been studied, especially in data-rich, highly dynamic, and adversarial environments. This project will also open up new possibilities to enhance U.S. Army s capability in future tactical battlefields by effectively and securely teaming companion robots with soldiers.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Oct 15, 2018
- Source ID
- W911NF1710447
Entities
People
- Chuan Yue
Organizations
- Army Contracting Command
- Colorado School of Mines
- United States Army