Otwarty dostęp

Analysis and synthesis of function data of human movement

Yunlong Ma
Yunlong Ma
 oraz 
Sanaa Sharaf
Sanaa Sharaf
   | 22 lis 2021
Applied Mathematics and Nonlinear Sciences's Cover Image
O artykule
Poprzedni artykuł
Następny artykuł
Zacytuj
Data publikacji: 22 lis 2021
Zakres stron: 917 - 926
Otrzymano: 17 cze 2021
Przyjęty: 24 wrz 2021
DOI: https://doi.org/10.2478/amns.2021.2.00086
Słowa kluczowe
© 2021 Yunlong Ma et al., published by Sciendo
This work is licensed under the Creative Commons Attribution 4.0 International License.

Fig. 1

Data curve during the walking exercise cycle.
Data curve during the walking exercise cycle.

Fig. 2

Derivative function graph.
Derivative function graph.

Fig. 3

Comparison of left foot data and waist data.
Comparison of left foot data and waist data.

Fig. 4

Period extraction results of a single measurement unit.
Period extraction results of a single measurement unit.

Fig. 5

Periodic extraction results.
Periodic extraction results.

Fig. 6

Comparison of algorithm results based on WARD. LPP, partially preserving projection; RP, random projection; WARD, wearable sensor behaviour recognition database.
Comparison of algorithm results based on WARD. LPP, partially preserving projection; RP, random projection; WARD, wearable sensor behaviour recognition database.

Fig. 7

Single sample data sequence.
Single sample data sequence.

Comparison of algorithm results based on WARD

Window size Feature extraction algorithm Classifier Average time spent on classification (s) Recognition rate (%)
Traditional algorithm 1 200 Time-frequency domain characteristics DT <1 92.6
Traditional algorithm 2 45 LPP DSC N/A 85.1
Traditional algorithm 3 40 RP SRC N/A 81.6
Algorithm 45 Periodic extraction of calibration starting point SVM <1 98.9

Behaviour description in WARD

Numbering Behaviour category Behaviour description
1 Walk normally Walk forward for >10 s
2 Walk counterclockwise Go counterclockwise for >10 s
3 Walk clockwise Clockwise for >10 s
4 To the left Turn left on the spot for >10 s
5 Turn right Turn right on the spot for >10 s
6 Go up the stairs Go up >10 stairs
7 Down the stairs Go down >10 stairs
8 Jogging Jogging lasts >10 s
9 Jump Jump in place >5 times
10 Push wheelchair Push the wheelchair for >10 s

Comparison of classification results of different classifiers

Classifier Recognition rate (%)
SVMs 98.9
DT 89.5
Naive Bayes 86.9
K-neighbours 93.9

Confusion matrix of 10 dynamic behaviour categories

1 2 3 4 5 6 7 8 9 10 Recognition rate (%)
1 97 0 0 0 0 1 0 0 0 2 97
2 0 98 0 1 0 0 0 0 0 2 98
3 0 0 100 0 1 0 1 0 0 0 100
4 0 0 0 100 0 0 0 0 0 0 100
5 0 0 0 0 98 0 0 0 0 0 98
6 0 0 0 0 0 98 1 1 1 0 98
7 0 0 0 0 0 2 94 1 0 3 94
8 0 0 0 0 0 0 0 99 0 0 99
9 0 0 0 0 0 1 5 0 100 0 100
10 1 0 0 0 0 0 0 0 1 100 100
Total recognition rate 98.4
eISSN:
2444-8656
Język:
Angielski
Częstotliwość wydawania:
Volume Open
Dziedziny czasopisma:
Life Sciences, other, Mathematics, Applied Mathematics, General Mathematics, Physics
Kanał RSS czasopisma