Boulic Model Lab Report
A human of height 1.8m moves with a constant speed of 0.9 m/s along the speci¯ed direct paths to radar inside the forest. The human is modeled using PEC ellipsoids, created based on the Boulic model as described in Section 2. The forest is modeled using identical trunks of 3 meters height and
11 cm radius which are distributed uniformly inside the 40£40m2 region. PEC is assumed for the trunk material properties as well as the ground. A trunk density of 3.8% (#/m2) is assumed, which amounts to 35 trunks within the half power beamwidth of the antenna, with average closest neighbor distance of 5 m. These forest parameters are chosen to re°ect the structure and spatial patterns of the trees in a typical old-growth forest [43].
The Doppler spectrogram …show more content…
from the scene is dependent on the path of motion with respect to the
radar.
To illustrate this e®ect, we consider two direct walking paths in the forest with di®erent starting points as depicted in Figure 12. Path 1 starts farther away from the radar, and should be exposed to more attenuation than Path 2. Also, as a consequence of being in the back of the forest, the human following Path 1 will be in the forward scattering zone of more trees than the human in Path 2. This is expected to result in more coupling e®ects from the trunks. We investigate the di®erent spectrograms for these paths in Figure 13. The same human starts with the same position and moves at the same speed for the two cases. The ¯rst step is taken with the left foot.
We observe that the motion can be detected for both scenarios, although the spectrograms di®er in nature as seen in the di®erent intensity levels for the left and right feet in Figure 13(a), while they are similar for Path 2 in Figure 13(b). The weaker signals for the left foot compared to the right foot in
Figure 13(a) are due to the coupling e®ects experienced by the human in Path 1. We also observe that the expected attenuation e®ects are visible, as the signature from the path closer to the radar (Path 2) is 20 dB higher than that of Path 1. This 20 dB di®erence corresponds to the round trip path loss for
the
10 meter separation between the two paths, based on the 1 dB/m speci¯c attenuation assumed for the forest at 5 GHz. Another feature we notice is that, despite the attenuation, the micro-Doppler signature still reveals the human motion 30m into the forest. Finally as expected, the static contribution from the scene is strong across the 0 Hz band for both cases, masking the lower speed motions due to the torso. We demonstrate in Figure 14, the spectrogram from the same scene for both paths without
Time (
11 cm radius which are distributed uniformly inside the 40£40m2 region. PEC is assumed for the trunk material properties as well as the ground. A trunk density of 3.8% (#/m2) is assumed, which amounts to 35 trunks within the half power beamwidth of the antenna, with average closest neighbor distance of 5 m. These forest parameters are chosen to re°ect the structure and spatial patterns of the trees in a typical old-growth forest [43].
The Doppler spectrogram …show more content…
from the scene is dependent on the path of motion with respect to the
radar.
To illustrate this e®ect, we consider two direct walking paths in the forest with di®erent starting points as depicted in Figure 12. Path 1 starts farther away from the radar, and should be exposed to more attenuation than Path 2. Also, as a consequence of being in the back of the forest, the human following Path 1 will be in the forward scattering zone of more trees than the human in Path 2. This is expected to result in more coupling e®ects from the trunks. We investigate the di®erent spectrograms for these paths in Figure 13. The same human starts with the same position and moves at the same speed for the two cases. The ¯rst step is taken with the left foot.
We observe that the motion can be detected for both scenarios, although the spectrograms di®er in nature as seen in the di®erent intensity levels for the left and right feet in Figure 13(a), while they are similar for Path 2 in Figure 13(b). The weaker signals for the left foot compared to the right foot in
Figure 13(a) are due to the coupling e®ects experienced by the human in Path 1. We also observe that the expected attenuation e®ects are visible, as the signature from the path closer to the radar (Path 2) is 20 dB higher than that of Path 1. This 20 dB di®erence corresponds to the round trip path loss for
the
10 meter separation between the two paths, based on the 1 dB/m speci¯c attenuation assumed for the forest at 5 GHz. Another feature we notice is that, despite the attenuation, the micro-Doppler signature still reveals the human motion 30m into the forest. Finally as expected, the static contribution from the scene is strong across the 0 Hz band for both cases, masking the lower speed motions due to the torso. We demonstrate in Figure 14, the spectrogram from the same scene for both paths without
Time (