Adaptive UAS Intent Estimation for Cyber-Physical Security
Sarocha Jetawatthana, Thanakorn Khamvilai · 2025
This research addresses the critical need for improved aircraft intent estimation to support the safety and security of modern aerospace and aviation systems. By leveraging machine-learning algorithms and dynamical theory, our research provides an adaptive estimation framework to increase the accuracy of UAV state and intent estimates. The framework allows for mismatch in the state estimation model by utilizing a neural network trained online to compensate for prediction errors. The neural network weights are adjusted by a Lyapunov-based update law. The resulting intent is derived from the marginalization of the covariance matrix.