Air Traffic Security: Aircraft Classification Using ADS-B Message’s Phase-Pattern

Abstract

Automatic Dependent Surveillance-Broadcast (ADS-B) is a surveillance system used in Air Traffic Control. With this system, the aircraft transmits their own information (identity, position, velocity, etc.) to any equipped listener for surveillance scope. The ADS-B is based on a very simple protocol and does not provide any kind of authentication and encryption, making it vulnerable to many types of cyber-attacks. In the paper, the use of the airplane/transmitter carrier phase is proposed as a feature to perform a classification of the aircraft and, therefore, distinguish legitimate messages from fake ones. The feature extraction process is described and a classification method is selected. Finally, a complete intruder detection algorithm is proposed and evaluated with real data.

Aircraft Mode S Transponder Fingerprinting for Intrusion Detection

“…An example is reported in Figure 3, where the phase patterns of consecutive messages are reported (the six phase features are extracted from these patterns). For this particular transponder, the phase pattern switches between three (almost interleaved) patterns: one quadratic and the other two broken-linear (as classified in [29]). It follows that, for this particular aircraft, the distributions for some features will be nonstationary (for short time observation), changing time-by-time or (for long-term observation) multimodal.…”

“…In some previous works (for example in [27,28]), the possibility to classify the aircraft Mode S transponders with respect to the time series of the transmitted messages or by the use of phase transient and carrier frequency are investigated. Moreover, in [29,30], the carrier phase pattern along the message is used for transponder classification by the use of Neural Networks or K-NN classifiers. In the latter, different approaches to extract the signature are also compared: single-message fingerprint or group-message fingerprint.…”

“…The carrier phase along the message. It can be estimated by estimating 112 different phase valueŝ φ m , one for each pulse of the ADS-B message using a Maximum Likelihood Estimator, as shown in [29]. The second-grade best-fitting polynomial, y = ax 2 + bx + c, of this sequence can be derived for each message.…”

“…In [29], the authors have already done a measurement campaign to evaluate carrier phase along messages as a transmitter feature, discovering that at least seven different types of transmitter are commonly used.…”

Detecting ADS-B Spoofing Attacks Using Deep Neural Networks

“…A malicious aircraft can also masquerade as another aircraft equipped with an ADS-B transponder of a similar implementation (e.g., from the same manufacturer) to avoid detection. In [14], Leonardi et al proposed to fingerprint aircraft through phase patterns Fig. 1.…”