MétaCan
Menu
← Back to cohort
Record W7141778074

Securing the ADS-B Protocol: A Comprehensive Framework Combining Bandwidth-Efficient Broadcast Authentication and Deep Learning-Based Intrusion Detection

2025· other· fr· W7141778074 on OpenAlexfundaboutno aff
Mikaëla Stéphanie Ngamboe Mvogo

Bibliographic record

VenuePolyPublie (École Polytechnique de Montréal) · 2025
Typeother
Languagefr
Field
Topic
Canadian institutionsnot available
FundersNatural Sciences and Engineering Research Council of CanadaConsortium de Recherche et d’innovation en Aérospatiale au Québec
KeywordsIntrusionImpossibilityThird party
DOInot available

Abstract

fetched live from OpenAlex

RÉSUMÉ: L’ADS-B (Automatic Dependent Surveillance–Broadcast) est un protocole de communication aéronautique, qui permet aux aéronefs de diffuser en continu leur position ainsi que d’autres données de navigation essentielles. Aujourd’hui largement déployé par les prestataires de services de navigation aérienne, tels que NAV CANADA, son utilisation est obligatoire dans plusieurs espaces aériens. Cependant, ce protocole présente une vulnérabilité majeure : l’absence de mécanismes d’authentification et de vérification de l’intégrité des messages transmis. Cette faille expose l’ADS-B à des cyberattaques, notamment via des radios logicielles, permettant à des acteurs malveillants d’émettre de faux messages, de rejouer des messages authentiques ou de brouiller les communications. Les conséquences potentielles d’une attaque sont graves : augmentation des risques pour la sûreté aérienne, perturbations opérationnelles et financières pour les compagnies aériennes (retards, annulations de vols, perte de confiance des passagers), ainsi que des enjeux géopoli-tiques liés à la sécurité et à la souveraineté des espaces aériens. Les travaux visant à sécuriser l’ADS-B ont révélé deux limites principales. D’une part, les schémas d’authentification proposés peinent à concilier les exigences de cybersécurité avec les contraintes opérationnelles imposées par la norme. En effet, la compatibilité descendante —indispensable pour que tous les récepteurs puissent décoder les messages — et les limites strictes de bande passante, qui imposent d’éviter toute augmentation du trafic sur un canal déjà saturé, rendent cette conciliation particulièrement difficile. D’autre part, bien que prometteurs, les systèmes de détection d’intrusion (SDI) fondés sur l’apprentissage profond montrent des limites face à des attaques sophistiquées ou inédites. Leur latence, inadaptée aux contraintes de temps réel de la gestion du trafic aérien, pose un défi supplémentaire. Pour répondre à cette problématique, cette thèse propose une approche intégrée combinant le protocole CABBA (Compatible Authenticated Bandwidth-efficient Broadcast for ADS-B) et un système de détection d’intrusions sophistiquées basé sur l’architecture xLSTM (extended long-short term memory). La solution CABBA repose sur le schéma d’authentification TESLA (Timed Efficient Stream Loss-tolerant Authentication), une modulation par superposition de phase et une infrastruc-ture à clé publique basée sur des certificats. Elle garantit l’authentification des émetteurs, l’intégrité des messages, la compatibilité avec les récepteurs existants et une efficacité spec-trale optimale. Afin de valider CABBA, nous l’avons implémenté et testé avec des radios logicielles. Nous avons mené des tests de compatibilité descendante avec des récepteurs ADS-B commerciaux et des équipements d’aviation générale. Nous avons également évalué le taux d’occupation du canal 1 090 MHz (pour les messages CABBA) et analysé le taux d’erreur binaire (BER, bit-error rate). Les résultats montrent que CABBA assure la compatibilité descendante, n’engendre qu’une surcharge minime du canal de communication et atteint un BER accept-able pour des valeurs de Eb/N0 supérieures à 14 dB. En complément, nous avons développé le premier système de détection d’intrusions pour l’ADS-B fondé sur une architecture xLSTM, optimisé par apprentissage par transfert. Par rapport à un SDI basé sur un transformeur, ce modèle atteint un score F1 de 98,9 % (contre 94,3 %) et s’avère efficace pour détecter les attaques graduelles, qui altèrent progressivement la conscience de la situation des pilotes et des contrôleurs aériens. Bien que sa latence d’inférence (7,26 s) soit supérieure à celle du transformeur (2,1 s), elle reste adaptée aux intervalles de rafraîchissement des radars secondaires (5 à 12 s). Ainsi, la solution CABBA et le système de détection d’intrusion basé sur l’architecture xL-STM constituent un cadre de sécurité complémentaire : CABBA assure la prévention grâce à l’authentification des diffusions, tandis que le système de détection pallie les risques résiduels liés aux limites de CABBA. ABSTRACT: Automatic Dependent Surveillance–Broadcast (ADS-B) is an aeronautical communication protocol that enables aircraft to periodically broadcast their position and other navigation data. By providing continuous and precise information about aircraft location, ADS-B func-tions as a surveillance technology employed by Air Navigation Service Providers (ANSP), such as NAV CANADA, to support Air Traffic Control (ATC) services. The use of ADS-B is mandated in many countries; however, it remains vulnerable to cyberattacks. The sys-tem lacks built-in mechanisms for authenticating messages or ensuring their integrity. Using readily available and inexpensive equipment, such as software-defined radios (SDR), mali-cious actors can inject false messages into the ADS-B system, replay legitimate messages, or jam communications. If left unaddressed, the exploitation of this vulnerability could com-promise the effectiveness of ADS-B and lead to wide-ranging consequences. These include heightened risks to aviation safety, operational and financial disruptions for airlines, such as delays, flight cancellations, and loss of passenger confidence, as well as broader geopolitical implications related to the security and sovereignty of national airspaces. Previous efforts to secure the ADS-B protocol have revealed significant limitations. Broadcast authentication schemes have struggled to balance security with operational requirements, such as backward compatibility and bandwidth efficiency. For data integrity verification, deep learning-based intrusion detection systems (IDS) have shown limited effectiveness against sophisticated or novel attacks, raising safety concerns. These concerns are often compounded by inference delays resulting from the complexity of modern deep-learning (DL) architectures. Given these challenges, the question of how to secure ADS-B while satisfying the technical and operational requirements that underpin its safety remains unresolved. To address this gap, we propose the Compatible Authenticated Bandwidth-efficient Broadcast protocol for ADS-B (CABBA), an enhanced broadcast authentication scheme that integrates Timed Efficient Stream Loss-tolerant Authentication (TESLA), phase-overlay modulation techniques, and certificate-based public key infrastructure (PKI). CABBA provides entity authentication, data origin authentication, and data integrity. To validate compliance with existing standards, we developed an SDR-based implementation and conducted backward compatibility tests with both commercial and general aviation ADS-B receivers. We also measured the 1 090 MHz channel occupancy rate following the ITU-R SM.2256-1 recommen-dation and performed a bit-error rate analysis. The results indicate that CABBA preserves backward compatibility, introduces negligible communication overhead, and achieves acceptable error rates for Eb/N0 values above 14 dB. Complementing this preventive solution, we designed the first extended long short-term mem-ory (xLSTM)-based intrusion detection system for ADS-B. Using a transfer learning strat-egy — pre-training on benign traffic and fine-tuning with tampered messages — the system was benchmarked against a transformer-based IDS. The xLSTM model demonstrated supe-rior performance, achieving an F1-score of 98.9 % compared to 94.3 % for the transformer. It also proved effective in detecting gradual attacks that progressively undermine situational awareness. Latency analysis showed that the 7.26-second inference time of the xLSTM model remains within the Secondary Surveillance Radar refresh interval (5–12 s), although it may pose constraints in time-critical scenarios, such as those in aircrew situational awareness and collision avoidance. The transformer-based IDS achieved a lower latency (2.1 s) but at the expense of reduced detection accuracy. Together, CABBA and the xLSTM-based IDS provide a complementary security framework: CABBA provides prevention through broadcast authentication, whereas the IDS mitigates the residual risks associated with the limitations of CABBA.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.

How this classification was reachedexpand

Full frame machine prediction

Teacher imitation

Not calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.

metaresearch head score (Codex)0.003
metaresearch head score (Gemma)0.003
Version: metacan-v3-hybrid-931329e0061cValidation status: machine_predicted_unvalidated
Candidate categoriesnone
Consensus categoriesnone
DomainCandidate signal: none · Consensus signal: none
Study designCandidate signal: Simulation or modeling · Consensus signal: Simulation or modeling
GenreCandidate signal: Empirical · Consensus signal: none
Teacher disagreement score0.006
Threshold uncertainty score0.014

Distilled classifier scores by category (both heads)

CategoryCodexGemma
Metaresearch0.0030.003
Meta-epidemiology (narrow)0.0010.001
Meta-epidemiology (broad)0.0010.001
Bibliometrics0.0010.000
Science and technology studies0.0010.001
Scholarly communication0.0020.002
Open science0.0020.003
Research integrity0.0020.002
Insufficient payload (model declined to judge)0.0020.001

Machine scores (provisional)

The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.

Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.

Opus teacher head0.008
GPT teacher head0.248
Teacher spread0.240 · how far apart the two teachers sit on this one work
Validation statusscore_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it

Classification

machine, unvalidated

Machine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.

The models applied no category: nothing in the taxonomy fit this work.
Study designSimulation or modeling
Domainnot available
GenreEmpirical

How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".

Quick stats

Citations0
Published2025
Admission routes2
Has abstractyes

Explore more

Same venuePolyPublie (École Polytechnique de Montréal)→French-language works237,207→