In this paper, we propose a low complexity algorithm to separate collided RFID signals. By carefully analyzing the collided signals when two tags with the same bit durations respond to a reader simultaneously, we observed that the individual amplitudes of each tag appear in pairs and the two tags' information bits are different, and regular RFID signals with larger amplitudes appear if the two tags' information bits are the same. Exploiting this feature, our algorithm can separate the collided signals of two tags to enhance the identification efficiency. Moreover, we implement our algorithm based on a software-defined reader platform to verify the validity of our algorithm. Simulation and experiment results show that the overall system throughput with our method is significantly increased in comparison with the conventional Gen-2 MAC scheme.