Utilizing Automated Robots to Recalibrate WiFi Fingerprint Maps for Indoor Location Estimation

WiFi fingerprinting is a localization technology that leverages existing wireless infrastructures in indoor settings. Although fingerprints can be calibrated to a high degree of accuracy, the precision of such maps changes due to environment changes. We propose a system that automatically re-calibrates the WiFi fingerprint maps using a mobile robot. As the robot traverses the building, it encounters RFID anchor tags embedded in the floor that helps the robot to determine its exact location. At these locations, the robot samples a WiFi fingerprint and associates it with the RFID tag location in a location database. Using this technique, we are able to overcome difficulties in fingerprinting calibration including human error and lack of samples. With this system we are able to attain twice the accuracy of state of the art systems: from 10 feet of accuracy to 3.5 feet.