This paper reports a novel system that real-timely detects the respiration and the heartbeat during sleep for the purpose of early diagnosis of Sleep Apnea Syndromes. Real-time discrimination of the signal components related to the respiration and the heartbeat was realized using the Independent Component Analysis from the observed signals of PVDF films, which are placed on the bed under the bed sheet. Body movement is taken into consideration in order to prevent the system from divergence and to improve the discrimination accuracy. Efficacy of the system is testified through experiments.