Team project / Physical-layer security
POLARIS
Exploring security at the level of the communication channel.
The project
A polar-code wiretap demonstrator built with six teammates. We explored how differences between a legitimate receiver’s and an eavesdropper’s channels affect decoding.
My contribution
I contributed to implementing and validating subchannel selection, binary-channel simulations and bit-error-rate analysis in Python.
Reported team result
The public case study reports a 200-frame binary symmetric channel simulation with Bob’s crossover probability at 0.05 and Eve’s at 0.25. At code length 1,024, list size 8 and 385 message subchannels, reported bit-error rates were 0.0035 for Bob and 0.2925 for Eve.
These are reported stochastic team results; the original seed and frame-level records are unavailable for exact reproduction. A decoding-error gap alone does not demonstrate end-to-end secrecy.
Recognition
The project was selected as one of two winning Semester 8 projects, with Thales mentorship. The work and its recognition belong to the team.
An opportunity to connect coding theory, simulation and a concrete engineering question.