Intelligent Writing System for Patients with Upper Limb Disabilities

Subjects with neurological problems and neurological disorders can be recovered in the dedicated recovery centers and at home. The paper presents a smart writing system that uses dedicated movements to recover specific problems of the hands. The system is based on a smart pen equipped with vibration...

Full description

Saved in:
Bibliographic Details
Published in:Journal of sensors Vol. 2022; pp. 1 - 18
Main Authors: Hnatiuc, Mihaela, Alpetri, Domnica, Arif, Muhamad, Vicoveanu, Dragos, Chiuchisan, Iuliana, Geman, Oana
Format: Journal Article
Language:English
Published: New York Hindawi 15.04.2022
John Wiley & Sons, Inc
Subjects:
Accelerometers
Algorithms
Bluetooth
Classification
Clustering
Data acquisition systems
Data analysis
Disabilities
Emotions
Experiments
Handwriting
Human-computer interaction
Machine learning
Neural networks
Neurological diseases
Physiology
Remote sensors
Sensors
Shape recognition
Signal processing
Software
Styli
Support vector machines
Systems design
Time series
Writing
ISSN:1687-725X, 1687-7268
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Subjects with neurological problems and neurological disorders can be recovered in the dedicated recovery centers and at home. The paper presents a smart writing system that uses dedicated movements to recover specific problems of the hands. The system is based on a smart pen equipped with vibration, acceleration, and gyroscope sensors that give remote data to a computer via an ESP Arduino board via Bluetooth. The purpose of the system design is to serve in applications such as recognizing the writing, drawings, or emotional and physiological states of the writer. The data is preprocessed by extracting the characteristics. They will be used in a prediction and classification system using machine learning (ML) algorithms. The paper proposes a new method of using the dynamic processing of certain geometric shapes that are recognized and analyzed to diagnose the mobility of the subjects’ hands. Using methods of clustering data as K-means and classification as support vector machine (SVM), the results obtained from the data analysis using the TensorFlow platform have an accuracy of over 70%. Due to the adaptation algorithms used, the system can be customized, learning the hand movement of the subject.
Bibliography:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:1687-725X
1687-7268
DOI:10.1155/2022/1005061