Determining mobile device indoor and outdoor location in various environments: Estimation of user context

To obtain a person´s location information with high accuracy in mobile device, it is necessary for a mobile device to switch its localization method depending on whether the user is indoors or outdoors. We propose a method to determine indoor and outdoor location using only the sensors on a mobile device. To obtain a decision with high accuracy for many devices, the method must consider individual difference between devices. We confirmed that using a majority decision method reduces the influence of individual device difference. Moreover, for highly accurate decisions in various environments, it is necessary to consider the differences in environments, such as large cities surrounded by high-rise buildings versus suburban areas. We measured classification features in different environments and the accuracy of classifier constructed using these features was 99.6%.