TY - GEN
T1 - Comparison of fingerprint matching methods for wi-fi indoor positioning
AU - Song, Haomin
AU - Jiang, Renjie
AU - Liu, Dawei
N1 - Publisher Copyright:
© 2018 IEEE.
PY - 2018/12
Y1 - 2018/12
N2 - Currently, positioning in indoor environment enjoys heavy demands in many public places. Considering that methods for outdoor positioning such as satellites localization are not applicable in indoor environment, Wi-Fi indoor positioning technology, regarding signal indicators as fingerprints, is a popular approach. However, most prior researches focus on the single device in position identification. As a result, the Euclidean distance similarity method, which is widely used for fingerprint matching of a single device, is not suitable for multiple devices. To solve the problem, this research aims to find a better matching method for multiple devices. An experiment is set to find the characteristics of Wi-Fi Received Signal Strength Indicator (RSSI) and create fingerprints for every position point. Since at least 92% RSSIs at one position point centralize at a specific value in the stable environment, the most centralized value is selected as fingerprints. Using the collected data and fingerprints, two matching methods - Euclidean Distance method and Cosine Similarity method, are compared in fingerprint matching for the same and different devices. The experiment results demonstrate that, by using two methods, the precision for same device matching is similar. However, for different device matching, Cosine similarity method, which has obviously increased the match accuracy, is a better method in fingerprint matching.
AB - Currently, positioning in indoor environment enjoys heavy demands in many public places. Considering that methods for outdoor positioning such as satellites localization are not applicable in indoor environment, Wi-Fi indoor positioning technology, regarding signal indicators as fingerprints, is a popular approach. However, most prior researches focus on the single device in position identification. As a result, the Euclidean distance similarity method, which is widely used for fingerprint matching of a single device, is not suitable for multiple devices. To solve the problem, this research aims to find a better matching method for multiple devices. An experiment is set to find the characteristics of Wi-Fi Received Signal Strength Indicator (RSSI) and create fingerprints for every position point. Since at least 92% RSSIs at one position point centralize at a specific value in the stable environment, the most centralized value is selected as fingerprints. Using the collected data and fingerprints, two matching methods - Euclidean Distance method and Cosine Similarity method, are compared in fingerprint matching for the same and different devices. The experiment results demonstrate that, by using two methods, the precision for same device matching is similar. However, for different device matching, Cosine similarity method, which has obviously increased the match accuracy, is a better method in fingerprint matching.
KW - Cosine similarity
KW - Euclidean distance
KW - Fingerprint
KW - Indoor positioning
KW - Wi-fi
UR - http://www.scopus.com/inward/record.url?scp=85070806519&partnerID=8YFLogxK
U2 - 10.1109/CompComm.2018.8780846
DO - 10.1109/CompComm.2018.8780846
M3 - Conference Proceeding
AN - SCOPUS:85070806519
T3 - 2018 IEEE 4th International Conference on Computer and Communications, ICCC 2018
SP - 748
EP - 752
BT - 2018 IEEE 4th International Conference on Computer and Communications, ICCC 2018
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 4th IEEE International Conference on Computer and Communications, ICCC 2018
Y2 - 7 December 2018 through 10 December 2018
ER -