With the declining birthrate and aging population, the proportion of single-living households, two-person households of elderly couples, and two-person households of an elderly and a child is increasing in Japan. In this situation, attention has been focused on the development of remote monitoring of the safety and security of these households. There are several techniques for the positioning system of people in residences and social welfare facilities. In the future, use of the gateway devices that aggregate wireless signals from various sensors will become popular with the spread of the Internet of things. These devices will make it possible to estimate the position of people who put on a wearable beacon using the received signal strength at the gateway devices. In this research, we propose a method using the support vector machine for the estimation, and compare the correct answer rates of estimation results with 2-dimensional, 3-dimensional and 4-dimensional support vector machines.