Sleep plays a significant role in human health. Along with the emergency of diseases related to sleep, sleep monitoring is becoming a hotspot. Traditionally, Polysomnography (PSG) is considered as the gold standard for sleep monitoring, but it is expensive, time-consuming, and uncomfortable, especially not ideal for long-term sleep studies. Thus, exploring alternative sleep monitoring method has ... [Show full abstract] drawn much attention of researchers. Currently most of the sleep monitoring methods require complicated signals and feature extraction process. However, this paper innovatively proposes a method using simple signal of heart rate and respiration rate, and integrating the prior knowledge of experts to simplify the feature extraction process, which can make the subsequent classifier distinguish wake state and sleep state easily. In this paper, we used multi-layer neural network as a classifier, getting a Cohen's kappa value of 0.739, and the overall accuracy of classifying sleep and wake state reached to 88%. Furthermore, we improved the ability of detecting wake states greatly by oversampling method, and the sensitivity and specificity reached to 91.3% and 82.3% respectively. Nowadays, the development of various bed sensors makes it easy to obtain the heart rate and respiration rate signals of human beings. Therefore, the method proposed in this paper is conducive to people's long-term sleep monitoring more conveniently and has great significance to people's health.