Acupuncture Manipulation Measurement and Identification Based on Miniature Multiaxis Force/Torque Sensor

Acupuncture manipulation is a key factor that directly affects efficacy, but it lacks accurate quantification and identification at present. For the current problems of low precision, complex operation, interference with operation, single identification category, and small sample size, it is the fir...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:IEEE sensors journal Ročník 24; číslo 20; s. 33806 - 33815
Hlavní autoři: Xu, Jingjing, Song, Aiguo, Chen, Yuhan
Médium: Journal Article
Jazyk:angličtina
Vydáno: New York IEEE 15.10.2024
The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
Témata:
Accuracy
Acupuncture
Acupuncture manipulation
Algorithms
Calibration
Classification
Error analysis
Force
force/torque (F/T) sensor
identification
Multiaxis
Performance evaluation
Sensor phenomena and characterization
Sensors
Support vector machines
support vector machines (SVMs)
Torque
Torquemeters
ISSN:1530-437X, 1558-1748
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:Acupuncture manipulation is a key factor that directly affects efficacy, but it lacks accurate quantification and identification at present. For the current problems of low precision, complex operation, interference with operation, single identification category, and small sample size, it is the first time that identification of four acupuncture manipulations is realized using a miniature multiaxis force/torque (F/T) sensor in this article. The sensor has the advantages of lightweight, high precision (the linear error is less than 1%), little interference with operation, and simple operation. First, a measurement system based on F/T sensor is constructed. Second, acupuncture manipulation data are collected using sensor and then processed by preprocessing and feature extraction. Finally, k-nearest neighbors (KNNs), support vector machines (SVMs), and random forest (RF) classification algorithms are selected for identification. The experiment results indicate that SVM classifier is more suitable for this system, and this model achieves an accuracy of 90.75% through fivefold cross validation. To evaluate detailed classification performance, validation is conducted on testing set. The accuracy of lifting-thrusting (LT) manipulation is 94.2%, and that of twirling-rotating (TR) manipulation is 85.8% in confusion matrix. This study has a good application prospect in acupuncture manipulation evaluation, training, and clinic.
Bibliografie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:1530-437X
1558-1748
DOI:10.1109/JSEN.2024.3453909