Aircraft Trajectory Estimation and Guidance Mode Identification using Bayesian Filtering Techniques

he dissertation explores advanced methods for aircraft trajectory estimation and trajectory prediction (TP), which are critical for the next generation of Air Traffic Management (ATM) systems operating under the Trajectory-Based Operations (TBO) paradigm. As ATM systems evolve, more responsibilities like flight plan adjustments and separation management are expected to shift from ground-based controllers to aircraft, emphasizing the need for precise trajectory estimation. 

The research primarily focuses on improving single-aircraft trajectory estimation and guidance mode identification, important for safe and efficient operations of both ground-based and airborne Decision Support Tools (DSTs). The core of the work is the development of sophisticated Interacting Multiple Model (IMM) algorithms enhanced with advanced filtering techniques, including Kalman filters and Sequential Monte Carlo (SMC) methods. These enhancements address challenges such as managing uncertainties, potential model mismatches, and identifying hidden guidance modes, all of which significantly impact TP. A dynamic model for trajectory estimation and TP requires various inputs, including measurement data, weather forecasts, aircraft performance models, and operational constraints. However, a key challenge is the availability of aircraft intent, which includes operational instructions and guidance modes that are required for accurately predicting trajectories. Identifying guidance modes is essential for improving TP performance, as it directly influences how the aircraft is controlled during flight. 

The main objectives of this dissertation are to characterize the estimation and guidance mode identification problem and enhance existing filtering methodologies using Bayesian techniques that rely on multiple-model approaches. The work proposes an optimal IMM approach that uses Kalman filter-based methods, particularly the Extended Kalman Filter (EKF), to manage the nonlinear dynamics predominant in vertical trajectory profiles. Results show that the maximum percentage error in aircraft mass estimation is minimal, and other state variables are also accurately estimated with low delays in tracking hybrid jumps and identifying guidance modes. 

Potential future directions include integrating lateral navigational maneuvers into system models, expanding the state vector to include additional flight parameters, and adapting these methodologies for multi-aircraft tracking scenarios. These enhancements aim to further refine TP capabilities, advancing ATM systems towards greater efficiency and safety.