Improving Physical Layer Secret Key Generation in Fast Fading Environments Using Prediction

Physical layer security (PLS) is an alternative to traditional information security mechanisms. Physical layer secret key generation (PSKG) is a PLS technique that uses channel randomness to generate keys for encryption algorithms. Although several works investigated PSKG in slow fading channels, fast fading scenarios remain a challenge. We propose two predictors to deal with such scenarios, one based on the Recursive Least Squares (RLS) and another based on the Dual Extended Kalman Filter (DEKF). The aim is to mitigate the channel variation effects during PSKG, reducing the errors between keys generated by legitimate users.