Air Data Sensor Anomaly Detection Using Spectral Signal Monitoring

The patent replaces basic mechanical threshold checking with spectral analysis (frequency domain processing) to detect sensor errors. By transforming pressure sensor signals into frequency spectra and analyzing spectral characteristics, the system can identify subtle errors that do not manifest as obvious range violations, thereby improving detection sensitivity while maintaining reliability.

The patent changes the parameter space for error detection by moving from time-domain threshold checking to frequency-domain spectral analysis. This parameter transformation enables the detection of subtle sensor errors through spectral signature comparison, allowing the system to identify anomalies that would be invisible to conventional range-based monitoring.

The patent implements a multi-functional processing system where the same spectral analysis infrastructure serves multiple purposes: baseline generation, real-time error detection, and sensor identification. By creating a universal processing framework that handles various air data sensors (pitot, static, angle of attack) through common spectral analysis routines, the system achieves high detection sensitivity without proportionally increasing complexity.

The patent performs preliminary spectral baseline generation during normal operation before errors occur. This pre-computed baseline serves as a reference for rapid error detection, allowing the system to compare current spectral signatures against historical norms without requiring complex real-time analysis of every parameter, thereby reducing ongoing processing complexity while maintaining high sensitivity.