GPS/Wireless LAN 기반 실내외 측위 알고리즘 개발 및 정확도 향상 연구

Abstract

본 논문은 GPS / Wireless LAN 기반 실내외 위치 추적 알고리즘 및 시스템개발에 관한 연구 이다. 이러한 연구 과제 목표를 위해 실내/외에서 Wi-Fi 수신 특성을 실험적으로 확인하였고 이를 토대로 여러 환경적 변수 요인 고려한 경로 손실 모델을수립하였다. 환경적인 요인에 따른 오차를 최소화하기 위한 측위 알고리즘의 성능 향상을 위해 Particle Filter의 측위 알고리즘 성능을 시뮬레이션을통해 확인하였다. 이를 통해 실외 지역, 건물인접 지역, 실내 지역에서 정확한 위치 추적이 가능한 GPS / WPS 하이브리드 측위 시스템을 Matlab 기반시뮬레이션을 실시 하였고 스마트폰 기반의 테스트베드를 구현함으로써 범용성과 실용 가능성을 높였다. GPS / WPS 하이브리드 시스템은 실내 또는 음영지역과 같은 GPS의 신호의 수신이 양호하지 않는 지역에서 위치 추정이 가능하게 하기 위해 GPS와 WPS를 결합하여 더 정확한 위치 정보를 사용자에게 제공 할 수 있었다. Seamless GPS / WPS 하이브리드 시스템을 개발하기 위해 파티클 필터를 이용한 위치 추정을 하였다. 건물 인접지역에서 GPS와 WPS의 위치 정보에 weighting factor를 이용한 Hybrid Positioning Engine을 적용하여 GPS Mode, Hybrid Mode, WPS Mode로의 변환을 자동적으로 변경되도록 설계하였다. 또한 주위 환경에 따라 수신 전파의 특성이 달라진다. 이러한 특징 요소를 각 위치 마다 적응력있게 적용하기 위한 알고리즘을 제안하였다. 또한 Weight Least square를 이용하여 다중경로로 인한 불안정한 Wi-Fi 신호의 영향을 감쇄시키는 알고리즘 그리고 과거의 입력 신호와 현재의 입력신호를 이용한 Mahalanobis distance를 이용해 outlier 신호를 판별하여 측위에 제외하여 정확도를 향상시키는 알고리즘을 제안하였다. 이들 제안한 알고리즘은 Matlab을 통해 가상 시뮬레이션과 실제 환경에 서 스마트폰을 이용하여 수집한 RSS 데이터를 이용하여 측위 테스트를 통해 실용성과 구현 현실성을 확인하였다.|Localization or Positioning is important in the field of wireless sensor networks that has attracted important research interest recently. The interest in positioning techniques is highly anticipated to grow the advances in the wireless communication techniques and the sensing techniques. There are many positioning techniques for indoor and outdoor environment. In outdoor settings, accurate Global Positioning System (GPS) based positioning has been accessible all over the world for almost a decade. Unfortunately, GPS is not available for some specific environments such as indoors, underpasses and tunnels. For these reasons, several technologies such as RFID, ZigBee, Bluetooth or Wireless Local Area Network (WLAN) have been recently proposed for indoor localization. Many researchers put their efforts into using existing wireless signal systems to determine location such as cellular phone systems, RFID, ZigBee, etc. RFID, ZigBee, and Bluetooth are not optimally suitable for indoor localization due to the maximum coverage and cost aspect. The maximum coverage of them is 2.5 10m at the most and the additional hardware is required to use these signal. Therefore, Wi-Fi based positioning system (WPS) emerged as an idea that can solve the positioning for in certain situations. Wi-Fi based positioning system (WPS) emerged as an idea that can solve the positioning in certain situations like indoors, underpasses and tunnels, taking advantage of the rapid growth of wireless access points in urban areas. In this thesis, I proposed a real-time tracking system using path loss model to convert distance between Mobile Station (MS) and Access Point (AP) from Received Signal Strength Indicators (RSSI). The position of the MS is tracked in real-time by particle filtering. The performance of the mobile tracking by particle ltering is demonstrated via MATLAB and iPhone real tracking test. The proposed various indoor/outdoor localization algorithm. Firstly, GPS/WPS hybrid positioning technique which give us estimation of the position indoors as well as outdoors. Secondly, I proposed variable path loss model to adapt optimal environment factors for indoor positioning system. Thirdly, noise reduction method in tough radio environment is proposed for accurate positioning and finally, I proposed the estimation method of the AP position. These algorithm improves the accuracy and effective of localization performance. The detailed explanation shows in the thesis.; Localization or Positioning is important in the field of wireless sensor networks that has attracted important research interest recently. The interest in positioning techniques is highly anticipated to grow the advances in the wireless communication techniques and the sensing techniques. There are many positioning techniques for indoor and outdoor environment. In outdoor settings, accurate Global Positioning System (GPS) based positioning has been accessible all over the world for almost a decade. Unfortunately, GPS is not available for some specific environments such as indoors, underpasses and tunnels. For these reasons, several technologies such as RFID, ZigBee, Bluetooth or Wireless Local Area Network (WLAN) have been recently proposed for indoor localization. Many researchers put their efforts into using existing wireless signal systems to determine location such as cellular phone systems, RFID, ZigBee, etc. RFID, ZigBee, and Bluetooth are not optimally suitable for indoor localization due to the maximum coverage and cost aspect. The maximum coverage of them is 2.5 10m at the most and the additional hardware is required to use these signal. Therefore, Wi-Fi based positioning system (WPS) emerged as an idea that can solve the positioning for in certain situations. Wi-Fi based positioning system (WPS) emerged as an idea that can solve the positioning in certain situations like indoors, underpasses and tunnels, taking advantage of the rapid growth of wireless access points in urban areas. In this thesis, I proposed a real-time tracking system using path loss model to convert distance between Mobile Station (MS) and Access Point (AP) from Received Signal Strength Indicators (RSSI). The position of the MS is tracked in real-time by particle filtering. The performance of the mobile tracking by particle ltering is demonstrated via MATLAB and iPhone real tracking test. The proposed various indoor/outdoor localization algorithm. Firstly, GPS/WPS hybrid positioning technique which give us estimation of the position indoors as well as outdoors. Secondly, I proposed variable path loss model to adapt optimal environment factors for indoor positioning system. Thirdly, noise reduction method in tough radio environment is proposed for accurate positioning and finally, I proposed the estimation method of the AP position. These algorithm improves the accuracy and effective of localization performance. The detailed explanation shows in the thesis.