An indoor positioning system is usually operated by WiFi-based TDOA Trilaterlation or Radio Map Fingerprint for locating positions. In this way, many WiFi signals in buildings can be easily detected. The TDOA Trilaterlation method has the advantages that with only more than 3 WiFi signals, it is possible to detect the user’s position. However, this method has high distance errors due to fluctuations in signal strength. On the other hand, the fingerprint method requires many signals to recognize places and shows less distance error compared to the TDOA method. Despite this, it needs a tremendous size of the search are enlarges. The ISI Lab. is researching an indoor positioning system based on Fingerprint method. We have designed to remedy Fingerprint method’s shortcoming of surveying radio map by increasing location accuracy.
The lab especially focuses on:
- decreasing error distance using sensor fusion
- devising new methods of radio map survey
- developing a location service platform
WiFi Radio Map of KAIST
Smartphones have many sensors like WiFi, Bluetooth, and gyro sensor. by using theses sensors together, accuracy of positioning systems can be improved. Incorporating these sensors is called “Sensor Fusion”. While this method makes some progress in developing higher levels of positioning systems, practically implementing this approach is difficult. Some of the signals of sensors have errors such as Bluetooth signal strength fluctuation due to obstacles from the sensors. Therefore, the key to sensor fusion is to minimize and correct these errors by using algorithms.
Collaborative localization is a widely used method for wireless sensor networks in GNSS (Global Navigation Satellite System) challenged environments which lack GPS signals. The concept of this system is that sensors share positioning related information within defined neighborhoods. As the more accurate the location information, the more precise the location based application becomes. Therefore, there is a trade-off between accuracy and the cost of location systems. This system is adventageous where a power saving locating system is required and only a tight budget is available.
Location Service Platform(KAILOS)
The ISI Lab. developed the indoor positioning platform KAILOS (Kaist Location System). This platform has been proven by being implemented in COEX indoor navigation system, “myCOEX”. Also, KAILOS proved its location accuracy by participating in the IPIN 2014 competition and showing the smallest distance error (5.7m) among 10 teams.
KAILOS system includes:
- Hybrid Positioning Engine(XPS)
- Walking Survey Radio Map Collector
- Web-based Map Construction Tool