Weighted Pattern Matching Algorithm for Indoor Location of Wireless Sensor Networks

How to effectively apply wireless sensor networks to achieve accurate location information in a complicated indoor environment is a challenging problem. In past years, varied indoor location algorithms have been proposed. The PM (Pattern Matching) mechanism is known as one of the most famous solutions for indoor location. Studies show that it is more accurate and stable than the other mechanisms. However, PM has drawbacks in the high initialization effort needed (higher training patterns). In this paper, we propose a Weighted PM (WPM) algorithm to reduce the initialization effort, further improve the accuracy and the locating stability of PM by weighting the healthiness of beacons to the matching function. Different from the conventional methods, which usually need additional equipment (such as infrared and ultrasound), the proposed WPM method is only based on the embedded RF chip to enhance the location accuracy in the indoor environment. Experiments show that WPM is outstanding over PM, with not only 50% fewer training patterns, but also 20.8% higher location accuracy. The stability of the location accuracy of WPM in indoor environment is up to 1.8 times greater than the others.