forked from wchill/HMP_Dataset
-
Notifications
You must be signed in to change notification settings - Fork 0
/
MANUAL.txt
123 lines (104 loc) · 4.21 KB
/
MANUAL.txt
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
Manual
------
1. Files tree
-----------------
ROOT
- brush_teeth (12 elements)
- climb_stairs (102 elements)
- climb_stairs_MODEL (20 elements)
- comb_hair (31 elements)
- descend_stairs (42 elements)
- drink_glass (100 elements)
- drink_glass_MODEL (20 elements)
- eat_meat (5 elements)
- eat_soup (3 elements)
- getup_bed (101 elements)
- getup_bed_MODEL (20 elements)
- liedown_bed (28 elements)
- pour_water (100 elements)
- pour_water_MODEL (20 elements)
- sitdown_chair (100 elements)
- sitdown_chair_MODEL (20 elements)
- standup_chair (102 elements)
- standup_chair_MODEL (20 elements)
- use_telephone (13 elements)
- walk (100 elements)
- walk_MODEL (20 elements)
- displayModel.m
- displayTrial.m
- MANUAL.TXT
- README.TXT
2. Human Motion Primitives
---------------------------
The dataset provides labelled recorded executions of a number of simple human
activities, which are defined as Human Motion Primitives (HMP):
1. brush_teeth: to brush one's teeth with a tootbrush
(complete gesture)
2. climb_stairs: to climb a number of steps of a staircase
3. comb_hair: to comb one's hair with a brush
(complete gesture)
4. descend_stairs: to descend a number of steps of a staircase
5. drink_glass: to pick a glass from a table, drink and put it
back on the table
6. eat_meat: to eat something using fork and knife
(complete gesture)
7. eat_soup: to eat something using a spoon
(complete gesture)
8. getup_bed: to get up from a lying position on a bed
9. liedown_bed: to lie down from a standing position on a bed
10.pour_water: to pick a bottle from a table, pour its content
in a glass on the table and put it back on the table
11.sitdown_chair: to sit down on a chair
12.standup_chair: to stand up from a chair
13.use_telephone: to place a telephone call using a fixed telephone
(complete gesture)
14.walk: to take a number of steps
3. Accelerometer specifications
--------------------------------
Type: tri-axial accelerometer
Measurement range: [- 1.5g; + 1.5g]
Sensitivity: 6 bits per axis
Output data rate: 32 Hz
Location: attached to the right wrist of the user with:
- x axis: pointing toward the hand
- y axis: pointing toward the left
- z axis: perpendicular to the plane of the hand
4. Data format
--------------
4.1. File naming convention
---------------------------
Each file in the dataset follows the following naming convention:
Accelerometer-[START_TIME]-[HMP]-[VOLUNTEER]
where:
- [START_TIME]: timestamp of the starting moment of the recording
in the format [YYYY-MM-DD-HH-MM-SS]
- [HMP]: name of the HMP performed in the recorded trial, following
the naming convention specified in Section 2 of this manual
- [VOLUNTEER]: identification code of the volunteer performing the recorded
motion in the format [gN] where:
- "g" indicates the gender of the volunteer
(m -> male, f -> female)
- "N" indicates the progressive number associated to the
volunteer
For example the file:
Accelerometer-2011-03-24-10-24-39-climb_stairs-f1.txt
refers to an accelerometer recording that was taken on March 24, 2011, starting from 10:24.39 a.m.. The recording refers to the HMP "climb_stairs" executed by the volunteer with ID "f1".
4.2 Acceleration data coding
----------------------------
Acceleration data recorded in the dataset are coded according to the following mapping:
[0; +63] = [-1.5g; +1.5g]
The conversion rule to extract the real acceleration value from the coded value is the following:
real_val = -1.5g + (coded_val/63)*3g
5. Dataset acquisition details
-------------------------------
5.1 Volunteers
--------------
The dataset is composed of the recordings of selected Human Motion Primitives perfomed by a total of 16 volunteers.
Basic information about the volunteers are reported in the table below:
| Gender| Age | Weight |
| M | F | Min | Max | Avg | Min | Max | Avg |
|-------|------------------|------------------|
| 11| 5 | 19 | 81 | 57.4 | 56 | 85 | 72.7 |
5.2 Hardware setup
------------------
The tri-axial accelerometer is embedded in an ad-hoc sensing device (40mm x 22 mm x 12 mm) that is attached at the right wrist of the user. Data transmission to the PC is wired, via a USB cable.