Indoor localization method based on AP and local linear regression algorithm

In this paper, an indoor localization method based on AP and local linear regression algorithm is presented. The method mainly makes use of Wi-Fi signal strength values collected from a small amount of deployed APs, and then the local linear regression algorithm is used to study and measure the degree of signal attenuation caused by the complex interior environment. So, the parameter of the algorithm is different for each AP. The circular area with radius of distance calculated by that algorithm with the signal strength value from deployed APs intersect, which forms a weighted overlapping area. By calculating the weight of each grid in this area, we can choose the one with the max probability as our predicted position. The experimental results show that the proposed method can control the positioning error in the range of 3 meters, which indicates that the method can provide a reasonable reference or even precise positioning result under low cost for indoor localization.