TY - GEN
T1 - Reducing Wi-Fi Fingerprint Collection Based on Affinity Propagation Clustering and WKNN Interpolation Algorithm
AU - Hu, Jiusong
AU - Liu, Hongli
AU - Liu, Dawei
AU - Yan, Zhi
AU - Xu, Kun
N1 - Publisher Copyright:
© 2018 IEEE.
PY - 2018/9/20
Y1 - 2018/9/20
N2 - The Wi-Fi-based indoor positioning system compares the online signal strength indicator (RSSI) and offline stored fingerprints to find the closest match to estimate the target location of the device. However, the problem is that the process of collecting fingerprints is very laborious, time-consuming and expensive. It is challenging to solve this problem. We proposed a method to reduce fingerprint collection based on APC (affinity propagation clustering) and WKNN (Weighted K-Nearest Neighbor) interpolation algorithm to solve the problem in this paper. We use the APC algorithm to break the known RPs (Reference points) into several clusters. The cluster representative is recorded. Then, we classify the unknown RPs as in the cluster which is physically nearest to cluster representative. Finally, in each cluster, we use fingerprints of known RPs with the WKNN interpolation algorithm to calculate fingerprints of the unknown RPs. We do experiments in a real environment. The results of real environmental experiments show that our method only needs 40% of the fingerprint to restore all the fingerprints in our environment, while the average localization accuracy only lost 8%. It only needs 70% of the fingerprints to restore all the fingerprints almost without any loss of average localization accuracy in our environment.
AB - The Wi-Fi-based indoor positioning system compares the online signal strength indicator (RSSI) and offline stored fingerprints to find the closest match to estimate the target location of the device. However, the problem is that the process of collecting fingerprints is very laborious, time-consuming and expensive. It is challenging to solve this problem. We proposed a method to reduce fingerprint collection based on APC (affinity propagation clustering) and WKNN (Weighted K-Nearest Neighbor) interpolation algorithm to solve the problem in this paper. We use the APC algorithm to break the known RPs (Reference points) into several clusters. The cluster representative is recorded. Then, we classify the unknown RPs as in the cluster which is physically nearest to cluster representative. Finally, in each cluster, we use fingerprints of known RPs with the WKNN interpolation algorithm to calculate fingerprints of the unknown RPs. We do experiments in a real environment. The results of real environmental experiments show that our method only needs 40% of the fingerprint to restore all the fingerprints in our environment, while the average localization accuracy only lost 8%. It only needs 70% of the fingerprints to restore all the fingerprints almost without any loss of average localization accuracy in our environment.
KW - WKNN
KW - Wi-Fi-based indoor positioning system
KW - affinity propagation
UR - http://www.scopus.com/inward/record.url?scp=85055674722&partnerID=8YFLogxK
U2 - 10.1109/IMCEC.2018.8469697
DO - 10.1109/IMCEC.2018.8469697
M3 - Conference Proceeding
AN - SCOPUS:85055674722
T3 - Proceedings of 2018 2nd IEEE Advanced Information Management, Communicates, Electronic and Automation Control Conference, IMCEC 2018
SP - 2463
EP - 2468
BT - Proceedings of 2018 2nd IEEE Advanced Information Management, Communicates, Electronic and Automation Control Conference, IMCEC 2018
A2 - Xu, Bing
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 2nd IEEE Advanced Information Management, Communicates, Electronic and Automation Control Conference, IMCEC 2018
Y2 - 25 May 2018 through 27 May 2018
ER -