STATE: Safe and Threatening Adversarial Trajectory Engine

NU 2025-140

INVENTORS

• Venkatramanan Subrahmanian* (McCormick School of Engineering, Computer Science)
• Chongyang Gao
• Lirika Sola Or Sholla
• Natalia Denisenko
• Tonmoay Deb
• Valerio La Gatta

STATE uses a GAN-based framework to generate synthetic safe and threatening drone trajectories. It helps security officials test and refine defense strategies in cities lacking prior flight data.

BACKGROUND

Cities face increasingly crowded airspace as drone usage expands for delivery, healthcare, real estate, and urban sensing. Existing tracking methods fall short due to high costs and limited data. This situation drives the need for simulation tools that support robust security planning.

STATE employs a conditional GAN architecture to generate drone trajectories based on designated threat levels. A pre-trained threat classifier guides the process, ensuring spatial realism and threat consistency. In tests with law enforcement experts, STATE achieved up to 75.8% improvement in trajectory plausibility and 35.8% in threat alignment compared to five baseline methods. This framework supports early warning assessments and enhances urban security simulations.

MARKET OPPORTUNITY

The global market for Unmanned Traffic Management (UTM), the enabling infrastructure for safe urban drone operations, is valued at $1.5 billion in 2025 and is projected to reach $5.8 billion by 2030, expanding at a remarkable CAGR of 31.0%. This growth is propelled by the rapid commercial adoption of drones for logistics, healthcare, and sensing, alongside the establishment of new regulatory frameworks for urban airspace. The primary market for this simulation technology includes municipal public safety agencies, urban planners, federal aviation authorities, and operators of critical infrastructure. (Source: MarketsandMarkets: "Counter-Drone Market - Global Forecast to 2030").

TRL-5 Prototype Validated in Relevant Environment: The prototype has been integrated with simulated urban airspace and evaluated by law enforcement experts.

• Synthetic trajectory generation: Produce safe and threatening drone flight paths for simulation.
• Security system calibration: Enhance threat assessment systems with realistic simulation data.

• Improves testing accuracy: Enhances defense planning with realistic trajectory simulations.
• Bridges data gaps: Generates synthetic data in regions lacking flight history.
• Enhances threat assessment: Achieves significant improvements in spatial plausibility and threat alignment.
• Cost effective: Reduces reliance on expensive and extensive real-world data collection.

• Venkatramanan Subrahmanian et al, "STATE: Safe and Threatening Adversarial Trajectory Engine". Preprint, 2025

US Patent Pending