Dataset for Indoor Positioning Using IEEE 802.11ac and 802.11ah (IPIN 2024)

This is the dataset for the paper "Performance Evaluation of Fingerprint-Based Indoor Positioning Using RSSI in 802.11ah" presented at IPIN 2024.

T. Matsunaga, I. Arai, Y. Atarashi, A. Endo and K. Fujikawa, "Performance Evaluation of Fingerprint-Based Indoor Positioning Using RSSI in 802.11ah," 2024 14th International Conference on Indoor Positioning and Indoor Navigation (IPIN), Kowloon, Hong Kong, 2024, pp. 1-7, doi: 10.1109/IPIN62893.2024.10786180.

Files

The correspondence between the datasets mentioned in the paper and the file names is as follows:

Mention in Paper	File Name
Full dataset	files/full_dataset.csv
Small dataset	files/small_dataset.csv

Item Names

Item Name	Description	Unit
index	Collection time (format: ISO 8601)	-
11ac_ap**_rssi	RSSI, Receive the beacon frame sent from the installed AP with ID:** to the receiver of the measurer.	-
11ah_receiver_**_rssi	RSSI, Receive the beacon frame sent from a measurer's 11ah AP to the installed 11ah receiver with ID:**	-
x	x-coorof the measured location	m
y	y-coordinate of the measured location	m
z	z-coordinate of the measured location	m
qw	Orientation at the time of measurement (quaternion qw)	-
qx	Orientation at the time of measurement (quaternion qx)	-
qy	Orientation at the time of measurement (quaternion qy)	-
qz	Orientation at the time of measurement (quaternion qz)	-
floor	Floor number within the building	-
measurement_point	ID to identify one measurement session. The same ID indicates the measurer continued walking.	-
room	ID to identify the room
block_x	x-coordinate with 1m precision (truncated)	m
block_y	y-coordinate with 1m precision (truncated)	m
block_z	z-coordinate with 1m precision (truncated)	m

Mapping of 11ac APs and 11ah receivers that were geographically close to each other

11ac	11ah
11ah_receiver_02_rssi	11ac_ap21_rssi
11ah_receiver_03_rssi	11ac_ap27_rssi
11ah_receiver_04_rssi	11ac_ap18_rssi
11ah_receiver_05_rssi	11ac_ap26_rssi
11ah_receiver_06_rssi	11ac_ap28_rssi
11ah_receiver_07_rssi	11ac_ap14_rssi
11ah_receiver_08_rssi	11ac_ap29_rssi
11ah_receiver_09_rssi	11ac_ap05_rssi
11ah_receiver_10_rssi	11ac_ap19_rssi
11ah_receiver_11_rssi	11ac_ap22_rssi
11ah_receiver_12_rssi	11ac_ap16_rssi
11ah_receiver_13_rssi	11ac_ap08_rssi
11ah_receiver_14_rssi	11ac_ap17_rssi
11ah_receiver_15_rssi	11ac_ap09_rssi
11ah_receiver_16_rssi	11ac_ap10_rssi
11ah_receiver_17_rssi	11ac_ap12_rssi
11ah_receiver_18_rssi	11ac_ap13_rssi
11ah_receiver_19_rssi	11ac_ap20_rssi
11ah_receiver_20_rssi	11ac_ap07_rssi
11ah_receiver_21_rssi	11ac_ap15_rssi
11ah_receiver_22_rssi	11ac_ap23_rssi
11ah_receiver_23_rssi	11ac_ap11_rssi

Correction Notice

After the publication of the paper, we discovered an error in the data and would like to provide a correction. We intended to use the RSSI values of 23 geographically proximate pairs of 11ac APs and 11ah receivers as features; however, due to a mistake in the dataset construction program, one pair was included in the dataset despite not being geographically close. In addition to the features listed in the table titled Mapping of 11ac APs and 11ah receivers that were geographically close to each other, we were also experimenting with 11ah_receiver_01_rssi and 11ac_ap25_rssi, but these two values are not geographically proximate. Therefore, the results of the paper were based on an incomplete dataset where 22 pairs were close, but one pair was not close.

When conducting experiments using the correct 22 pairs, the results are as follows:

• Area Classification (Inside/Outside Furnace Room) Accuracy
  • Number of 11ac APs with highest F1 Score: 19 → 18
• Area Classification (Room) Accuracy
  • Highest F1 Score: 0.997 → 0.995
  • Number of 11ah receivers with highest F1 Score: 20 → 18
• ECDF
  • 11ac
    • Probability of positioning error less than 10m (%): 99.6 → 99.5
  • 11ah
    • Probability of positioning error less than 10m (%): 97.4 → 97.0
• Positioning Error (MedAE)
  • 11ac
    • Lowest MedAE: 2.537 → 2.564
    • Number of APs with lowest MedAE: 21 → 19
  • 11ah
    • Lowest MedAE: 3.722 → 3.745
    • Number of Receivers with lowest MedAE: 22 → 20