Calibration-free localizations and applications on U-care cloud

The key technique of the location-based service (LBS) is localization which is a kind of techniques for determining the location of mobile users (MU). One of the most common techniques for the location estimation of mobile users (MU) is the radio frequency (RF) site survey. The main idea of the technique is to build a signal strength model, called a radio map, in the off-line phase and to estimate the location of an MU by finding the best match from the radio map in the on-line phase. However, when signal strength values vary frequently due to the characteristics of environmental dynamics, the radio map will quickly become outdated, and recalibration requires considerable manual effort. A good positioning technique should be able to adapt to a dynamically changing environment. In this paper, the authors describe the design and implementation of a positioning system which can provide low-cost, but highly adaptable and precise positioning in the context of changeable radio environments. Instead of constructing a radio map, the authors use reference points (RP) installed in the localization area to continuously monitor changes in the environment. The authors then employ the minimum mean square error (MMSE) method to estimate the location of the MU. Experimental results show that the average error distance is about 3 meters. The positioning system has been implemented as a subsystem of the U-care project in the Chang Gung Health and Culture Village. From the results of questionnaire made by residents, the authors find that the most satisfaction system in the U-care project is the positioning and rescuing system because it can locate the exact position of a resident in need of emergency.