WiFi-based activity recognition dataset ////////////////////////////// WiAR dataset : WiFi-based Activity Recognition dataset ////////////////////////////
The WiAR dataset contains sixteen activities including coarse-grained activity and gestures performed by ten volunteers with 30 times every volunteer. Now, we only list three volunteers' activity dataset such as volunteer A, volunteer B, volunteer C. Of course, you can extract RSSI and CSI from raw data in the WiAR. We also provide some codes to beginners. The goal is to share new ideas with other reseachers to develop the field.
Section 1:Device Configuration
- T400 laptop, 5300 card.
- Sample rate: 30Hz
- Three antennae 具体参考：http://dhalperi.github.io/linux-80211n-csitool/installation.html my blog:http://blog.csdn.net/guolinlin11
Section 2: Data format
We provide raw data received by the receiver. The details are shown in following.
-
each activity has 30 samples.
-
ten volunteers participate in collecting activity data.
-
csi_ai_j.dat represent the jth sample of the ith volunteer.
Section 3: activities
There are sixteen activities as follows.
- horizontal_arm wave
- high arm wave
- two hands wave
- high throw
- drwa_x
- draw tick
- toss paper
- forward kick
- side kick
- bend
- hand clap
- walk
- phone call
- drink water
- sit down
- squat
these data is raw data. Therefore, user need to preprocess phase.
Section 4: codes(matlab)
- read_bf_file.m
The function lists several details including timestamp low, bfee_count, Nrx, Ntx, rssi_a, rssi_b, rssi_c, noise, perm, agc, rate, csi.
- get_csi_from_file The function obtains CSI. website:http://dhalperi.github.io/linux-80211n-csitool/faq.html
If your paper relate to my works, please cite our papers as follows.
@inproceedings{guolinlin_1, title={A Survey on motion detection using WiFi signals}, author={Linlin Guo, Lei Wang, Jialin Liu, and Wei Zhou}, booktitle={Proc. of IEEE MSN}, year={2016} }
@inproceedings{guolinlin_2, title={A Novel benchmark on human activity recognition using WiFi signals}, author={Linlin Guo, Lei Wang, Jialin Liu, Wei Zhou, Bingxian Lu}, booktitle={Proc. of IEEE Healthcom}, year={2017} }
@article{guolinlin_3, author={Linlin Guo, Lei Wang, Jialin Liu, Wei Zhou, and Bingxian Lu}, journal={Wireless Communications and Mobile Computing}, title={HuAc: Human Activity Recognition Using Crowdsourced WiFi Signals and Skeleton Data}, year={2018}, doi:={10.1155/2018/6163475} }
Finished paper and unsubmited:
[1] WiAR: An Public Dataset for WiFi-based Activity Recognition
[2] CAR: CSI-based Coarse-grained Activity Recognition
[3] A Survey on Approaches of Activity Recognition Using WiFi Signals
IF you need to our help, please contact with us by email. Email:[email protected].