Satellite manoeuvres are a major source of uncertainty in orbit prediction and can significantly reduce prediction accuracy if not properly identified. This thesis presents a framework for detecting satellite manoeuvres using only publicly available Two-Line Element (TLE) data within the German Aerospace Center (DLR) Improving Satellite Orbit Prediction Accuracy (ISOPA) framework. The proposed methodology analyses TLE consistency through residuals in the Radial–Transverse–Normal (RTN) frame, combining physically motivated anomaly scores with normalization, whitening, and machine-learning refinements. Beyond manoeuvre detection, the framework enables the characterization of satellite manoeuvring behaviour. Finally, detected manoeuvres are incorporated into the orbit-prediction process by excluding manoeuvre-contaminated data from model training, leading to significant improvements in long-term prediction accuracy. The results demonstrate that TLE consistency analysis provides a practical and effective basis for manoeuvre-aware orbit prediction in Space Situational Awareness applications. ... Read more