Key Takeaways
- A Morgan State University student team won a $25,000 IEEE Aerospace and Electronic Systems Society award for a lightweight onboard AI system that detects and responds to cyber‑attacks on UAVs without relying on ground links.
- The solution continuously learns normal sensor and control behavior, distinguishes benign anomalies from malicious tampering, and initiates safe autonomous actions such as subsystem isolation, sensor switching, or graceful degradation.
- By embedding security directly on the aircraft, the project addresses a critical weakness of conventional approaches that depend on uninterrupted communications for monitoring and analysis.
- The award requires the team to expand the scope, collaborate with an industry partner, and release the software publicly, turning the innovation into a shared resource for the broader aerospace‑cybersecurity community.
- Faculty advisers from Morgan’s Data Engineering and Predictive Analytics Lab and the Center for Equitable Artificial Intelligence and Machine Learning Systems guide the work, highlighting the university’s growing expertise in AI‑enabled cybersecurity for autonomous flight.
Project Overview and Award Recognition
Morgan State University’s Data Engineering and Predictive Analytics (DEPA) Research Lab garnered international acclaim when students Awotwi Baffoe and David Nyarko secured a $25,000 award from the IEEE Aerospace and Electronic Systems Society’s Cybersecurity Challenge Phase 2. Their proposal, titled “Lightweight Onboard AI for Cyber‑Resilient UAV Anomaly Detection and Safe Mission Response,” was chosen for its focus on secure and resilient avionics operation. The recognition places the Morgan team among a select group of innovators tackling one of aerospace cybersecurity’s most pressing challenges: protecting increasingly autonomous unmanned aerial vehicles from sophisticated digital attacks when ground‑based communications may be compromised or unavailable.
Why Onboard Security Is Essential for UAVs
Traditional security architectures often offload data to ground stations or cloud platforms for monitoring and analysis, assuming a reliable communications link will persist throughout a flight. However, an adversary capable of jamming or spoofing those links can simultaneously attack the aircraft’s sensors, navigation signals, or control systems, leaving the UAV blind and unable to seek help. At altitude, a compromised drone cannot “pull over” or reboot; it must instantly discern trustworthy data from malicious interference and act to preserve safety. This reality transforms cybersecurity from an IT concern into a core safety issue for autonomous flight.
Core Innovation: Lightweight Onboard AI Model
To address this vulnerability, the Morgan team designed a compact artificial intelligence model that runs directly on a UAV’s onboard computer, respecting the stringent size, weight, and power constraints of flight hardware. The model continuously learns the baseline patterns of normal behavior across the aircraft’s sensor suite and control systems by observing data streams during routine operation. By establishing a dynamic profile of expected activity, the AI can detect deviations that fall outside normal variations caused by turbulence, wear, or benign sensor noise.
Threat Detection and Anomaly Classification
When the onboard AI observes a departure from learned norms, it employs statistical and machine‑learning techniques to assess whether the anomaly likely stems from ordinary flight conditions or indicative of a cyber‑physical attack. The system evaluates factors such as the magnitude, timing, and correlation of sensor deviations, distinguishing between harmless fluctuations and signatures of tampering—such as sudden GPS spoofing, inertial sensor corruption, or unauthorized command injections. This real‑time classification enables the UAV to react swiftly before an attacker can exploit the vulnerability.
Autonomous Safe‑Response Mechanisms
Detection alone is insufficient; the system is programmed to initiate a safe response without awaiting instructions that may never arrive. Depending on the context and severity of the threat, the UAV can isolate the affected subsystem, switch to a trusted or redundant sensor source, transition to a reduced‑capability flight mode, or execute a pre‑planned safe recovery or landing maneuver. By degrading gracefully rather than failing outright, the aircraft preserves mission integrity and protects people and property on the ground, even when the link to the operator is severed.
Interdisciplinary Collaboration and Faculty Guidance
The project exemplifies the applied, interdisciplinary learning fostered at Morgan State University. Students Awotwi Baffoe and David Nyarko integrate knowledge from data engineering, machine learning, cybersecurity, and electrical and computer engineering to solve a complex real‑world problem. Faculty advisers—Kelechi Nwachukwu (adjunct faculty and engineering manager for DEPA), Peter Taiwo, D.Eng. (CEAMLS faculty researcher and DEPA lecturer), and Kofi Nyarko, D.Eng. (professor of Electrical and Computer Engineering, director of CEAMLS and DEPA)—provide mentorship, ensuring the work aligns with both academic rigor and practical aerospace requirements.
Award Conditions, Timeline, and Open‑Source Commitment
Morgan submitted the proposal on May 15, 2026, and received notification of selection on June 19, 2026, from Southwest Research Institute, which administers the IEEE AESS challenge. As part of the award, the review committee asked the team to broaden the project’s scope to address two challenge problems, collaborate with an industry partner, and publish the resulting software in a public repository. The team accepted these conditions and will complete the work during a three‑month performance period, with grant coordination handled through the Georgia Tech Research Institute. Making the code openly available will allow researchers worldwide to examine, test, and build upon the Morgan‑developed solution, amplifying its impact beyond the competition.
Broader Impact and Future Milestones
The recognition advances Morgan’s research profile in AI‑enabled cybersecurity for aerospace and unmanned systems, while also illustrating how student innovation can translate theory into tangible societal benefit. David Nyarko’s prior involvement with Team AgroVision—which earned a $50,000 award for a modular hydroponic farming system—demonstrates the breadth of research opportunities available to Morgan students and their ability to apply technical expertise across domains ranging from food systems to autonomous flight. The next milestone for Baffoe, Nyarko, and their advisers is to validate that a small, efficient layer of onboard intelligence can reliably enable an aircraft to recognize danger, protect itself, and chart a safe path forward—even when ground communications fall silent—thereby contributing a critical building block for the future of secure autonomous aviation.