The paper discusses the detection of Primary User Emulation Attacks (PUEA) in cognitive radio networks using a technique called Signal Activity Pattern Acquisition and Reconstruction System (SPARS). To improve the detection capability, a new method called Weighted Sequential Probability Ratio Test (WSPRT) is introduced, which enhances the true positive rate suffered by the SPARS model due to Byzantine failure. The study outlines the effectiveness of the proposed method through simulations, showing improvements in detecting smart attackers.