Detection of Pseudo Random Noise Generator's parameters for link analysis

Abstract

This paper presents the method how the parameters of Pseudo Random Noise Generator (PRNG) can be detected. The parameters to be detected include the primitive polynomial and the initial seed of the PRNG. Different test vectors of scrambled sequences are generated by changing the parameters of Linear Feedback Shift Register (LFSR). The test vectors are then passed one by one through the detection algorithm which calculates the weight difference for a range of primitive trinomials. Subsequently, the weight difference is calculated between the alphabets over GF(2) of decoded periods of LFSR The maximum weight difference thus obtained for specific polynomial depicts the most probable primitive trinomial. The initial seed of LFSR is then found by fixing the trinomial and calculating the weight difference for a range of seeds. The correct initial seed is depicted by the maximum weight difference.