Secure and Safe Assured Autonomy (S2A2)

NASA University Leadership Initiative Project (2020-2025)

Project Description

Aviation's future is likely to feature a diverse range of Advanced Air Mobility (AAM) systems, including Unmanned Aerial Systems (UASs) for cargo and delivery, personal air vehicles, and commercial Urban Air Mobility (UAM) Aerial Vehicles (AVs). However, there are significant challenges that could delay or even prevent these developments. To overcome these obstacles, research is strongly needed in various fields to leverage technologies such as autonomy, cybersecurity, Air Traffic Management (ATM), multi-redundant flight system architectures, and advanced wireless connectivity like 5G/6G.

Under the guidance and support from NASA, this University Leadership Initiative (ULI) project is to create new technologies and innovative operational concepts which will ensure safe, secure and robust integration of autonomous vehicles into AAM-tailored transportation infrastructure. All this must be done while maintaining inter-operability with current civil air transportation systems and associated safety standards.

This collaborative project with three academic partners and three industry partners is structured into four Technical Challenge (TC) areas, each designed to offer unique UAM solutions and provide a transition road-map for industry and government to utilize the research outcomes.

TC1 (Safe Perception, Coordination, Planning, and Navigation): The incorporation of low-altitude UAVs and UAM systems introduces new safety risks and requires reliable sensing, coordination, planning, and navigation to handle dynamic environments. TC1 focuses on integrating various algorithms and software components to address technological challenges. These include operating at different timescales and unifying diverse data types into a format compatible with current and future embedded flight control systems for UAVs and UAM systems.
TC2 (Secured Autonomy): For future secured autonomy in UAM systems, resolving issues surrounding operations in potentially hostile environments with highly dynamic situations is crucial. The system's inherent heterogeneity among cyber and physical components, its large scale, and the limited computation/processing capability of individual components may pose complex challenges. TC2 specifically investigates the system's capability to handle malicious cyberattack impacts, such as Denial-of-Service (DoS), False-Data-Injection (FDI), and stealthy attacks. The ultimate goal of the TC2 group is to develop vehicle-level defense strategies that can enhance the safety/security of UAM AVs despite cyber threats during their operations.
TC3 (Verification & Validation): The hybrid dynamics of complex cyber-physical systems within UAM networks are challenging to model and analyze for human interaction. Thus, scalable and modular V&V and T&E tools that can adapt to evolving specifications are needed. Effective V&V and T&E frameworks in TC3 offer provable correctness guarantees, enhance safety, and provide certification methodologies for autonomous airspace operations.
TC4 (System Integration): Integrating emerging secure and safe autonomy algorithms into practical flight software is challenging. Testing these algorithms in various environments—dynamic simulation systems, small-scale Indoor Flight Facilities (IFF), and large-scale Outdoor Flight Facilities (OFF)—and performing different missions is essential for robust integration and public trust. Industry partners play a crucial role in TC4.

Our FD&C/HS Lab has been involved in the TC2 research group. We are specifically focusing on developing numerous cyber threat management algorithms, including 1) Cyberattack Mitigation and Attenuation, which reduces the impact of cyberattacks; 2) Risk Assessment under Potential Cyberattacks, which can proactively quantify how much potential risk (e.g., crashes or collisions) can be obtained from cyberattacks. These research efforts are essential to enhance the level of safety and security of UAM AVs even though they encounter cyberattacks during their operations.

Research Directions and Outcomes

1) Cyberattacks Mitigation & Attenuation

Observer-Based Attack Resilient Controller Design

In this category, we have developed an observer-based resilient control strategy for multi-agent systems (MASs) when agents (e.g., UAM Aerial Vehicles) encounter False-Data-Injection (FDI) attacks in their sensors. One of the system vulnerabilities of a MAS is an attack propagation to the entire network of the MAS. In other words, the impact of cyberattacks can significantly degrade the performance of collaborative mission of MAS, such as formation control and velocity-matching consensus. To handle this problem, we have proposed a vehicle-level observer-based resilient control strategy based on the Lyapunov stability criterion. The estimated FDI attack signals are then integrated to the nominal controller (e.g., PID or state feedback) of agents such that the impact of FDI attacks can be directly mitigated. This greatly enhances the system resiliency even though the operation of MAS experiences sensor failures posed by FDI attacks.

Publications

S. Hwang, M. Cho, and I. Hwang, "Resilient Tracking Control Strategy for Discrete-Time Networked Control Systems under False-Data-Injection Attacks," Submitted to IEEE Transactions on Control of Networked Systems.

Figure: Graphical overview of the proposed distributed resilient control/optimization scheme against DoS attacks. Our proposed method allows UAM AVs to establish enhanced communication network by employing the network centroid control, so that the system's resiliency against DoS attacks can be increased.

Distributed Resilient Control and Optimization

In this category, we have developed a distributed resilient control and optimization strategy tailored for Urban Air Mobility (UAM) Aerial Vehicles (AVs). UAM systems are characterized by the heterogeneity of participating AVs. Participating AVs are expected to cooperate with each other while maintaining flexibility in individual missions and reacting to the possibility of cyberattacks and security threats. To address this security issue, we focus on the vulnerabilities of the UAM system against Denial-of-Service (DoS) attacks. We develop a distributed resilient control strategy for the AVs navigating through the UAM airspace to mitigate the effect of DoS attacks. A graphtheoretic vulnerability metric is proposed. Each AV can compute its vulnerability against DOS attacks in a fully distributed manner using this metric. Based on this computed metric, the AVs self-organize to minimize collision risk in the operating airspace after assessing self-vulnerability. This reconfiguration is also carried out in a fully distributed manner. The proposed control scheme is proven to reduce vulnerability in a probabilistic manner. This reduced vulnerability holds against DoS attacks with a known attack budget.

Publications

S. G. Clarke, S. Hwang, O. Thapliyal, and I. Hwang, “Distributed Denial-of-Service Resilient Control for Urban Air Mobility Applications,” Journal of Aerospace Information Systems, vol. 20, no. 12, pp.873-889, September 2023, doi.org/10.2514/1.I011222.

S. G. Clarke, O. Thapliyal, S. Hwang, and I. Hwang, "Attack-resilient distributed optimization-based control of multi-agent systems with dual interaction networks". AIAA SciTech Forum, San Diego, CA, Jan. 2022, doi.org/10.2514/6.2022-2342.

2) Risk Assessment under Potential Cyberattacks

Figure: MATLAB simulation with 4 agents under stealthy attacks. The blue ellipse represents the fleet-level security metric and the red ellipse shows the agent-level security metric.

Lyapunov-Based Reachability Analaysis

In this category, we have developed a risk assessment tool for multi-agent systems (MASs) in the face of cyberattacks (e.g., stealthy attacks and Denial-of-Service (DoS) attacks). Since the operations of MASs are strongly dependent on the communication between agents (i.e., UAM aerial vehicles), the impact of cyberattacks may lead to numerous detrimental situations, such as crashes and collisions. To address this cybersecurity issue for MASs, we have proposed an ellipsoidal risk assessment tool through the Lyapunov-based reachability analysis. This tool indicates how much potential threat can be expected from cyberattacks. We utilize this indicator as a security metric for the operations of MASs, including fleet and agent levels. This information plays a significant role in enhancing the safety/security of MASs despite threats from potential cyberattacks. Finally, we demonstrate the efficacy of the proposed method through the realistic Urban Air Mobility (UAM) common scenario, which has been proposed under the guidance from NASA.

Publications

S. Hwang, M. Cho, S. Kim and I. Hwang, "An LMI-Based Risk Assessment of Leader–Follower Multi-Agent System Under Stealthy Cyberattacks," in IEEE Control Systems Letters (L-CSS option with IEEE CDC 2023), vol. 7, pp. 2419-2424, June. 2023, doi: 10.1109/LCSYS.2023.3286775.

M. Cho, S. Hwang, and I. Hwang, "Risk Assessment of Multi-Agent System Under Denial-of-Service Cyberattacks Using Reachable Set Synthesis," Accepted in 2024 American Control Conference (ACC), Toronto, Canada, July 10-12, 2024.

Publications

S. Hwang, M. Cho, and I. Hwang, "Resilient Tracking Control Strategy for Discrete-Time Networked Control Systems under False-Data-Injection Attacks," Submitted to IEEE Transactions on Control of Networked Systems, June. 2024.
M. Cho, S. Hwang, and I. Hwang, "Risk Assessment of Multi-Agent System Under Denial-of-Service Cyberattacks Using Reachable Set Synthesis," Accepted in 2024 American Control Conference (ACC), Toronto, Canada, July 10-12, 2024.
S. Hwang, M. Cho, S. Kim and I. Hwang, "An LMI-Based Risk Assessment of Leader–Follower Multi-Agent System Under Stealthy Cyberattacks," in IEEE Control Systems Letters (L-CSS option with IEEE CDC 2023), vol. 7, pp. 2419-2424, June. 2023, doi: 10.1109/LCSYS.2023.3286775.
O. Thapliyal, and I. Hwang. "Data-driven Cyberattack Synthesis against Network Control Systems," IFAC-PapersOnLine, vol. 56, no. 2. pp. 8357-8362, 2023, doi.org/10.1016/j.ifacol.2023.10.1027.
S. G. Clarke, S. Hwang, O. Thapliyal, and I. Hwang, “Distributed Denial-of-Service Resilient Control for Urban Air Mobility Applications,” Journal of Aerospace Information Systems, vol. 20, no. 12, pp. 873-889, September 2023, doi.org/10.2514/1.I011222.
S. G. Clarke, O. Thapliyal, S. Hwang, and I. Hwang, "Attack-resilient distributed optimization-based control of multi-agent systems with dual interaction networks". AIAA SciTech Forum, San Diego, CA, Jan. 2022, doi.org/10.2514/6.2022-2342.

People

Sounghwan (Eric) Hwang, Ph.D. student [August 2021 ~ Current]
Sungsoo Kim, M.S. [Graduated in May 2024]
Omanshu Thapliyal, Ph.D. [Graduated in May 2023]
Shanelle G. Clarke, Ph.D. [Graduated in December 2022]

Media & News