Seamless Navigation Methodology optimized for Indoor/Outdoor Detection Based on WIFI

Smartphones with multiple sensors become more popular in contemporary society. The LBS (Location-Based Service), based on smartphones, develops rapidly. One of the most important prerequisite for LBS is to determine the indoor or outdoor environment and the location of the users, so that different service information can be offered. A seamless navigation methodology optimized for indoor/outdoor detection based on WIFI is proposed in this paper. The RSSI (Received Signal Strength Indication) values of WIFI collected by the smartphone are used by the AdaBoost algorithm to train the weak classifier into the strong classifier to distinguish the indoor/outdoor environment, as well as improving the overall accuracy of the indoor and outdoor seamless navigation. The classic AdaBoost method is improved in this paper: First, the RSSI value of a pair of APs (Access Points) is compared to construct a weak classifier, in order to solve the problem of device heterogeneity; Second, the optimized AdaBoost method adding several weak classifiers in each training phase, so that the abnormal conditions of APs can be reduced. When the device detects that the current environment is indoors or outdoors, the navigation mode is adapted to the current environment and the navigation accuracy can be improved. The experiment results indicate that the accuracy of indoor/outdoor detection is more than 97%, and it can significantly improve the continuity and the accuracy of the indoor/outdoor seamless navigation.