As robots increasingly need to communicate and work together, their vulnerability to cyberattacks grows. Our project has made significant strides in securing robotic networks by developing collaborative methods for detecting and preventing unauthorized actions.
Our methods combine the robots' cyber-physical capabilities—communication, sensing, and movement—enabling them to monitor each other's actions and identify attempts by an infiltrated robot to breach restricted areas (what we term a "deviation attack"). If a robot deviates from its expected path, the group detects and contains the threat, preserving overall network integrity.
Our contributions are organized across three main research areas. First, we developed a formal framework to define "deviation attacks" and potential attacker behaviors, providing a foundation for rigorous proofs of our methods' effectiveness.
Second, we created novel algorithms for robotic coordination that allow groups of robots to plan their paths and monitor each other’s activities, ensuring mutual accountability. Our algorithms apply to increasingly complex settings, from simple grid layouts to continuous-motion environments resembling real-world scenarios, such as environmental monitoring. These solutions can also suggest additional safeguards—such as deploying more robots for enhanced monitoring—when necessary.
Finally, we equipped robots with control tools for adaptive, on-the-fly threat response. By allowing robots to autonomously handle unforeseen obstacles, disturbances, and additional tasks, our approach minimizes reliance on centralized control, enhancing system flexibility. As a result, the robots can operate effectively even in changing environments, while still providing robust protections against deviation attacks.
Beyond these technical achievements, the project has had broader impacts in research, education, and community outreach. We provided valuable training to six graduate students, two undergraduates, and two high school students at Boston University and Northeastern University, giving them hands-on experience in cybersecurity and robotics. The skills they gained will be essential as they enter the workforce to address challenges in autonomous systems and cybersecurity.
To maximize the benefit of this research, we have also shared our results widely. Through seven academic papers, two PhD dissertations, and the development of an open-source software toolbox, our findings are accessible to researchers and practitioners worldwide, supporting further advancements in robotic security.

Last Modified: 11/06/2024
Modified by: Roberto Tron