case study

Detecting Runway Threats Before They Happen

NAVAIR, the Naval Air Systems Command, is the backbone of U.S. naval aviation. They design, build, test, and sustain the aircraft and systems that keep the Navy and Marine Corps mission ready, including the F-35 Joint Program Office, which manages the most advanced and expensive fighter jet program in Department of Defense history.

KUNGFU.AI partnered with the F-35 Joint Program Office to create a state-of-the-art AI system that detects and mitigates Foreign Object Debris (FOD) on airfields by analyzing radar data, environmental sensors, and aircraft telemetry. The solution significantly improved detection accuracy, reduced manual inspection efforts, and enabled edge deployment for real-time, low-bandwidth operation.

‍

AI Solution(s)

Computer Vision

Edge AI Deployment

Industry

Aerospace

Sweeping the Runway with Smarter Signals

Vision

Protecting Fleet Readiness With Smarter FOD Detection

Strengthen the F-35 Joint Program’s mission readiness by applying AI to reduce the risk of Foreign Object Debris (FOD) incidents that threaten aircraft safety and availability. The goal was to replace time-intensive manual inspections with a smarter, scalable system that proactively identifies risks across complex airfield environments.

Challenge

Modernizing a High-Stakes, Manual Safety Process

FOD poses an existential risk to aircraft engines—one ingestion can result in a $30M loss, including engine replacement, labor, and downtime. Traditional detection relied on manual foot patrols or costly radar systems, while the data required to modernize this process was difficult to access, fragmented, and proprietary.

The F-35 Joint Program Office (JPO)—the DoD’s largest and most complex acquisition program—oversees every aspect of the F-35 fighter jet for multiple military branches and international partners. Ensuring the aircraft’s operational readiness requires not only maintaining the jets themselves but also optimizing the infrastructure, systems, and procedures that support them.

Breakthrough

Transforming Sensor Noise Into Actionable Insight

What began as a feasibility test to improve a low-cost radar sensor evolved into a comprehensive analytics program. The team integrated a wide range of data sources, including radar imagery, environmental conditions, ATC (air traffic control) transcripts converted from audio to text to track aircraft movement, and direct telemetry from the F-35’s Inlet Debris Monitoring System (IDMS), which records when foreign material passes through the engine intake. To make sense of this complex data ecosystem, the team pioneered:

A novel patch-based computer vision approach that analyzes radar images by breaking them into smaller sections for improved detection accuracy
Synthetic data generation techniques that allowed the team to simulate FOD events and train models more effectively despite limited real-world data
An edge compute deployment strategy that miniaturized the model and enabled real-time analysis directly on radar equipment, eliminating the need for high-bandwidth infrastructure

These technical innovations not only improved detection sensitivity by up to 60% but also made the system generalizable enough to be deployed across other F-35 Joint Program airfields.

Outcome

A Scalable AI Platform That Elevates Safety and Readiness

The work yielded a new AI-based detection platform capable of proactively identifying high-risk FOD zones and guiding runway mitigation protocols. By reducing the risk of engine damage, the system enhances pilot safety, safeguards aircraft integrity, and strengthens overall fleet readiness—helping protect one of the Department of Defense’s most critical and costly investments. The solution is now being deployed across additional airfields, with low compute costs and significant operational impact.