BB84 Covert Channel

A covert channel embedded in the classical sifting phase of the BB84 Quantum Key Distribution protocol. The channel transmits hidden data by intentionally misreporting basis announcements, hiding the covert payload within the hardware's expected QBER noise floor.

Paper: Covert Channel via Error-Injection by Classical Misreporting in the BB84 Sifting Phase (link TBD)

Overview

This project extends an existing BB84 QKD simulation by Dhruv Bhatnagar with a covert channel implementation and statistical analysis tools.

Requirements

• Python 3.10+
• NumPy
• SciPy
• Matplotlib

Project Structure

Base BB84 Simulation (by Dhruv Bhatnagar)

File	Description
qkd_bb84_base.py	Core classes for Alice, Bob, quantum channel, and classical channel
qkd_experiment_base.py	Base experiment class with build, run, key generation, and validation phases
qkd_bb84_noiseless.py	Noiseless BB84 experiment driver
qkd_noise_model.py	Noisy channel and Hadamard gate failure models
qkd_eavesdropping_2.py	Intercept-resend eavesdropping simulation

Covert Channel Implementation (by Albie Snyder)

File	Description
covert_channel.py	CovertStateMachine class — rolling trigger mechanism, dual PRNGs, keystream
qkd_covert_alice.py	Covert Alice subclass — basis misreporting with message preamble and padding
qkd_covert_bob.py	Covert Bob subclass — covert bit extraction and preamble parsing
covert_experiment.py	Noiseless covert channel experiment driver
covert_noise_experiment.py	Noisy covert channel experiment driver
baseline_noise_experiment.py	Baseline noisy experiment (no covert channel) for comparison
statistical_analysis.py	KS test analysis for QBER and inter-error distance detectability
generate_figures.py	Generates all paper figures from saved experimental data
covert_test.py	Unit test for CovertStateMachine synchronization

Data and Figures

Directory	Description
Data/	Saved NumPy arrays from 1000-trial experiments (.npy files)
Figures/	Generated SVG and PNG figures for the paper

Quick Start

Run a basic covert channel demo (noiseless)

python covert_experiment.py

This transmits a 24-bit test message through the covert channel with trigger length k=7 and prints the recovered message.

Run a noisy covert channel experiment

python covert_noise_experiment.py

Same as above but with 3% channel noise simulating depolarizing errors.

Run the full statistical analysis

python statistical_analysis.py <k>

where <k> is the trigger length (e.g., 7, 8, 9, 10). Runs 1000 baseline and 1000 covert trials, performs KS tests on QBER and inter-error distance distributions, and saves results to .npy files.

Generate figures

python generate_figures.py

Loads saved data from Data/ and generates all paper figures to Figures/.

Key Parameters

Parameter	Description	Default
s_length	Number of qubits per session	65,536
trigger_length (k)	Length of trigger sequence	7
noise_error_rate	Channel depolarizing error rate	0.03
PSK	Pre-Shared Key (string)	"topsecret"

License

The base BB84 simulation is by Dhruv Bhatnagar. The covert channel extension is by Albert Snyder.