论文部分内容阅读
Crowdsourcing-based localization signal (e.g., wireless signal and environmental signal) database updating is becoming a trend for low-cost location-based service (LBS) and location-enhanced internet of things (L-IoT) applications. This paper proposes a crowdsourcing-based localization framework that updates both wireless (e.g., WiFi) and environmental (e.g., magnetic) signal maps autonomously. Compared to the existing indoor localization methods that do not involve GNSS, this paper introduces GNSS data as constraints for the optimization of crowd-sourced dead-reckoning (DR) solutions. Furthermore, to alleviate the issue that multiple localization signals do not interact, this paper combines multiple signals to a multi-source localization database, which is convenient and effective to be used to enhance localization. Through the use of the generated multi-source database for localization, there was an accuracy improvement of approximately 20 % compared to the state-of-the-art DR/Wireless/Magnetic integration solution.