A Feasibility Study on Smart Mattresses to Improve Sleep Quality

Abstract

Good sleep quality is essential, especially for clinical users. Sleep disorders not only impair the success rate of treatment but also delay recovery. They can seriously interfere with treatment outcomes and even endanger a user's life. In this study, we created a smart mattress containing 10 × 18 air packs and control units. Each air pack contains a set of pressure and height sensors and two air valves. Each row control unit can detect and adjust the pressure and height of each air bag in the row. When the bed body is turned on, it automatically initializes, adjusts the state of each air bag to the same height and pressure, and enters a slow scanning state. When perceived objects or people are lying on the bed, the bed automatically perceives the human body structure and body pressure matrix, increases the scanning speed for more timely and accurate measurements of the digital matrix and forming pressure by matrix-normalized processing, and then uses local pressure variance detection to automatically adjust to the sleeping position of the human body and thus achieve a uniform force distribution and a comfortable state. Finally, pressure matrix binarization was used to match sleeping position templates to identify the best template for automatic recognition of the sleeping position. The experimental results show that the sleeping position recognition method has high accuracy, recall, and precision. Our mattress is designed with interfaces for external devices. In future research, the smart mattress can connect to an auxiliary part of a smart ecosystem consisting of a smart pill box, a smart lighting system, and a microclimate system, which is expected to yield a more comprehensive intelligent ward to explore the possibility of improving sleep quality.