Research and implementation of an indoor positioning algorithm

Abstract: The goal of the Internet of Things’ sensing technology is to provide LBS(location-based services); a key technology is finding out how to positioning the sensing devices. For positioning outdoors, mature tech-nology such as GPS and cellular network location can be used. There is little research about indoor positioning, and there is no finished product on the market. This paper shows how to use both Wi-Fi and ZigBee signal for position-ing; Wi-Fi to find the area position and ZigBee to find the coordinate position. The main contribution of this paper is described in the follow-ing: This paper will present an algorithm using kNN on a Wi-Fi signal, as a way to find the location area of users. The GPS signal cannot be used indoors, but there are usually numerous Wi-Fi signals, that can be used for indoor positioning. In this design, to build a dataset containing the number of locations and the Wi-Fi signal strength list of each location. When indoor positioning is needed, the KNN algorithm is used to compare the user’s Wi-Fi signal strength with the dataset and find the location number. When precise positioning is needed, the ZigBee signal should be used. In this paper two different methods for precise positioning in are used, one is an improved algorithm of triangle centroid algorithm where the positioning accuracy depends on the number of anchor points and the interval of each point. The other method is the neural network method. This method could give stable result with only four anchor points. Finally, there is a comparison of the methods mentioned in this paper : the Wi-Fi fingerprint method, the ZigBee triangle centroid algorithm, and neural network method.