Reduced search space for the optimal design of linear phase FIR filter

Branch and Bound Technique (BBT) is an optimal method for digital filter design in the discrete space. However, this algorithm is increasingly complex, as it requires an exhaustive search of the optimal solution in the discrete space. BBT may only be used for filter length N≤8. The aim of this paper is to improve the convergence of the BBT algorithm. Therefore, we propose a novel method, which efficiently localizes the reduced region that contains the optimal solution. The proposed method tends to reduce the algorithmic complexity of BBT, preserving the optimality of the method. In this paper, we also schow, the performance of the method on few examples.