Movement detection for power-efficient smartphone WLAN localization

Mobile phone services based on the location of a user have increased in popularity and importance, particularly with the proliferation of feature-rich smartphones. One major obstacle to the widespread use of location-based services is the limited battery life of these mobile devices and the high power costs of many existing approaches. We demonstrate the effectiveness of a localization strategy that performs full localization only when it detects a user has finished moving. We characterize the power use of a smartphone, then verify our strategy using models of long-term walk behavior, recorded data, and device implementation. For the same sample period, our movement-informed strategy reduces power consumption compared to existing approaches by more than 80% with an impact on accuracy of less than 5%. This difference can help achieve the goal of near-continuous localization on mobile devices.