Computation of the k-error linear complexity of binary sequences with period 2n

Abstract

The k-error linear complexity(k-LC) of sequences is a very natural and useful generalization of the linear complexity(LC) which has been conveniently used as a measure of unpredictability of pseudorandom sequences, i.e., difficulty in recovering more of a sequence from a short, captured segment. However the effective method for computing the k-LC has been known only for binary sequences with period 2n (Stamp and Martin, 1993). This paper gives an alternative derivation of the Stamp-Martin algorithm. Our method can compute not only k-LC but also an error vector with Hamming weight ≤ k which gives the k-LC.