Wi-MIR: A CSI Dataset for wifi based multi-person Interaction Recognition
THE SHIN LAB @ KOREA UNIV
About Wi-MIR Dataset
The developed Wi-MIR dataset is pivotal to Wi-Fi-based human activity recognition, specifically for multi-person interactions (MIPs) recognition. When humans move in between the transmitting and receiving antennas, the movement of various body parts in between Wi-Fi signals’ propagation path generates changes in the signal reflections and refraction, which is evident by the channel state information (CSI) variations.
This dataset is collected within an indoor setting, specifically a furnished room measuring 8m x 6.1m, consisting of seventeen multi-person interactions (MPIs) conducted by eleven human pairs utilizing Wi-Fi. Each pair of subjects performed twenty trials of each of the seventeen MPIs, and the total number of 3,740 trials is recorded in our dataset (i.e., 17 interactions × 11 pairs of human × 20 trials). The publicly available CSI tool records the Wi-Fi signals transmitted from a Wi-Fi router (NETGEAR Nighthawk-R7000) operating as an AP to a desktop computer equipped with an intel 5300 Network Interface Card (NIC) operating as a receiver. The AP and receiver have three transmitting and receiving antennas, respectively and operate at a 5 GHz frequency band, specifically on channel number 36, with a channel bandwidth of 20 MHz. The recorded Wi-Fi signals consist of the Received Signal Strength Indicator (RSSI) and the Channel State Information (CSI) values. Our dataset can be exploited to advance Wi-Fi-based human activity recognition in different aspects, such as machine learning and Deep learning algorithms to recognize different MPIs.
The seventeen MPIs are below
Approaching
Bowing
Conversation
Departing
Exchanging objects
Handshaking
Helping standup
Helping walk
Hugging
Kicking with the left leg
Kicking with the right leg
Pointing with the left hand
Pointing with the right hand
Punching with the left hand
Punching with the right hand
Pushing
Touching another person