3.2. Data Collection

In order to measure CSI, we use two TPLINK N750 routers (two Atheros AR9580 Wi-Fi chipsets respectively loaded on the two industrial computers). One is transmitter (Tx) and the other is the receiver (Rx). The receiver router is equipped with three antennas with an interval of 0.1m, while the transmitter router is equipped with only one antenna. The bandwidth is 40 MHz, and the Wi-Fi frequency is 5 GHz. So, there are 114 subcarriers for each Tx-Rx pair. The sampling rate is 1000 packet/s.

We collect data in two indoor environments. In the first environment, we marked three positions on which the volunteers stand and perform designated actions. In the second environment, two locations are marked. As presented in Figure 2, the interval between the transmitter and the receiver is 2.6 m and the volunteer stands in the designated position facing the acquisition device. The volunteer sees the instructions on a screen placed in front of him/her. The details of the two indoor environments are shown as follows.

The layout of the two indoor environments, the upper one is Room1 and the lower is Room2.

The Room1 is a 10 × 13 m room. There are three desks with the size of 5 × 2 m in Room1. There are three windows on the right side of the room. The transmitter(Tx) and receiver(Rx) are separated by 2.6 m.

The Room2 is a 10 × 4 m room. There are two desks in Room2, the size of the desk on the left is 3 × 1 m and the size of right desk is 7 × 2 m. There is a window on the right side of the room. The distance between transmitter(Tx) and receiver(Rx) is 2.6 m.

The five volunteers stand at five different positions (three in room1 and two in room2) and complete the six designated gestures. Each gesture is repeated 20 times. So, we collected 3000 (5 volunteers × 5 positions × 6 gestures × 20 instances) instances. Each instance has four labels, which include the identity of volunteers, room number, location number, and action category.

We collect six common gestures, which are similar to the gestures in ARIL. These gestures include hand left, hand right, hand up, hand down, hand circle, and draw zigzag (from Widar3.0). These six gestures are presented in Figure 3. Please note that these gestures are commonly used in the field of human-computer interaction. The first four gestures have a well-defined directionality. In contrast, the last two gestures are more complex. We use an automated collection platform to collect the CSI information containing the specified gesture. The platform is developed by a member of our team. When collecting data, we place a monitor opposite to the volunteer to display the action acquisition instructions. A server is placed behind the monitor to provide the computing services and monitor the operations of the system. The volunteers stand at the designed location and face the monitor. Before data collection, we need to present the action category and collection times. After the operations of the system starts, the display requests the volunteer to raise the hand upwards. Please note that the preparation time is 3 s. After 3S, the volunteers start performing the action, and the system starts to collect data packets. The duration of data collection is 1.8 s. Afterwards, the display stops for 2S, and the volunteers can have a short break. Please note that the preparation time and rest time after collection can be adjusted.

The six common gestures, including hand up, hand down, hand left, hand right, hand circle, and draw zigzag.