AXIOM: A Zero-Trust Cyber-Physical Integrity Framework for Commercial EFB-to-FMS Data Pipelines

Modern commercial aircraft rely on Electronic Flight Bags (EFBs) to supply critical takeoff performance parameters, specifically (Takeoff Decision Speed), (Rotation Speed), and (Takeoff Safety Speed), directly to the Flight Management System (FMS). The implicit trust placed in this data pipeline represents an unaddressed cybersecurity gap: a confirmed DEF CON 31 disclosure demonstrated that the Airbus Flysmart+ EFB contained a disabled SSL validation control, permitting man-in-the-middle modification of flight-critical SQLite databases [1]. This paper presents AXIOM (Authenticated eXecution Integrity for Onboard Mission-critical computations), a three-layer middleware architecture that forces every EFB output to prove its physical validity before FMS execution is permitted. Layer 1 applies ECDSA P-384 cryptographic attestation against an Immutable Baseline Database (IBD) housed in a hardware Secure Enclave. Layer 2 applies real-time kinematic physics cross-validation using fourth-order Runge-Kutta (RK4) integration and vectorized Monte Carlo analysis (N = 1,000,000). Layer 3 performs continuous Synthetic Vision System (SVS) coherence monitoring with Exponential Moving Average (EMA) sensor fusion. Software-in-the-Loop (SITL) integration with the JSBSim A320 flight dynamics model supplies live aircraft mass and air density to Layer 2, and live altitude data to Layer 3. A test suite of 105 automated tests verified all components. End-to-end pipeline latency measures 1.825 ms (mean) on x86 hardware, projecting to 3.650 ms on ARM-based EFB hardware. Monte Carlo analysis over one million trials confirms at confidence for nominal A320 parameters. Layer 3 achieves a zero false alarm rate over 3,600 simulated approach timesteps under realistic AR(1) correlated sensor noise. All four tested attack scenarios (including a physically consistent, cryptographically valid payload signed with the trusted key) are detected and blocked.