Sleep monitoring systems that can be used in daily life for the assessment of personal health and early detection of diseases are needed. We are developing a system for unconstrained measurement of the respiration and heartbeat of a person on a soft rubber-based tactile sensor sheet. To extract faint heartbeat signals from pressure detected by the tactile sensor, improvement of the S/N ratio is needed. This process takes some time, and can be conducted at only a limited number of locations on the sensor. The suitable locations for the heartbeat detection depend on not only the location but also the lying posture of the person on the sensor. We propose a method for detecting the lying location and posture using a pattern recognition technique based on machine learning. In this paper, we describe the method and report the experimental results.