Phantom: Prediction of Human Motion with Distributed Body Sensors - Prediction and Classification of Respiratory Motion - page 52

Table 3.2 Characteristics of the motion data
Motion data Motion type Speed (cm/sec) Recording time (sec)

Chest_1 Slow motion 0.64-0.87 20.72

Chest_2 Moderate motion 6.7-0-7.91 20.40

Chest_3 Violent motion 24.84-32.63 22.18

Head_1 Slow motion 0.63-1.08 20.36

Head_2 Moderate motion 6.62-8.37 20.40

Head_3 Violent motion 16.07-67.70 21.43

Upper Body_1 Slow motion 0.68-1.48 20.83

Upper Body_2 Moderate motion 3.64-28.08 20.64

Upper Body_3 Violent motion 38.48-118.18 21.03

dataset) for each motion. We increased the group number (G) of EM process from
two to seven to show all the potential hyper-parameters. Please note that we have
eight sensory channels, so that we can show all available group numbers in Fig. 3.8 .
Figure 3.8 shows the hyper-parameters (b y , where y is a group number)
described in Eq. ( 3.4 ), based on the target motion data and group number (G).

β 1 β 2 β 3 β 4 β 5 β 6 β 7

0.5

Chest

0

2 3 4 5 6 7

Group number (G)

0.5 Head

0

2 3 4 5 6 7

Group number (G)

0.5 Upper Body

0

2 3 4 5 6 7

Group number (G)

Fig. 3.8 Hyper-parameter values based on the target motion data and group number