*3.2. Measurement Scenarios*

To verify the proposed human recognition scheme, we designed eight cases as scenarios to be carried out on the test-bed. A photo of each case is presented in Figure 11. In this paper, we measure the radar signal for 60 s in every case. That is, in each case, measurements were conducted for 60 frames. However, in each case, the number of the extracted feature vectors is 52, because of the initial data collection time of 8 s, as described in Section 2.3. 

– **Figure 11.** (**a**–**h**) Test scenarios represented as eight cases in a vehicle.

– In this paper, we only consider humans as living creatures. That is, we do not discuss other living creatures such as companion animals, because the recognition of the passenger as a human is most important. Moreover, we considered the seats in the vehicle, a box on a seat, and the vehicle body itself as inanimate objects. A detailed description of the scenario is given below.


For the measurements of cases #2–#5, we conducted the experiments with three males. The characteristics of these participants are given below. The photos in Figure 11 show human #2. In the experimental results in Sections 3.3 and 3.4, we present the results from human #2.

• Human #1 is 27 years old, 183 cm tall, and weighs 78 kg.


The CW Doppler radar operates at the 2.45 GHz ISM band, which is freely used for the purpose of industrial, science, and medical applications. The transmitted power of the radar is less than 5 dBm, which meets the regulation for human bodies. All the subjects joining in the experiment agreed with protocol, procedure, and process in the measurement.
