Design and application research of a flexible array plantar sensor based on P(VDF-TrFE)/SnO2NPS/GR for Parkinson’s disease diagnosis

• Published in: Polymer-plastics technology and engineering Vol. 63; no. 14; pp. 1975 - 1999
• Main Authors: Luo, Yi, Su, Peinan, Wu, Ying, Zhao, Zhidong
• Format: Journal Article
• Language: English
• Published: New York Taylor & Francis Ltd 21.09.2024
• Subjects: Classification; Diagnosis; Electrical impedance; Fourier transforms; Graphene; Infrared spectroscopy; Machine learning; Mechanical properties; Medical diagnosis; Open circuit voltage; Parkinson's disease; Piezoelectric films; Raman spectroscopy; Sensor arrays; Sensors; Short circuit currents; Spectrum analysis; Tin dioxide; Vinylidene; Vinylidene fluoride
• ISSN: 0360-2559; 1525-6111
• DOI: 10.1080/25740881.2024.2365278

This paper introduces a flexible array Plantar sensor fabricated through the high-voltage electrospinning process of poly(vinylidene fluoride-trifluoroethylene) (P(VDF-TrFE)) ，incorporating stannic oxide nanoparticles （SnO2NPS）and graphene (GR) composite nanofilm. The surface morphology, β-phase crystal content, piezoelectric performance, composition, and structure of the composite piezoelectric films were evaluated using Scanning Electron Microscopy (SEM), X-ray Diffraction (XRD), Fourier Transform Infrared Spectroscopy (FTIR), Raman Spectroscopy, and a vibration platform. Experimental findings reveal that P(VDF-TrFE)/SnO2NPS/GR composite films containing 5% SnO2NPS and 0.1% GR exhibit superior open-circuit voltage and short-circuit current, measuring 22.43 V and 12.95 uA, respectively. These values are approximately 1.59 times and 1.34 times higher than those of the 5% P(VDF-TrFE) composite film and about 2.37 and 2.16 times higher than those of pure P(VDF-TrFE). A flexible piezoelectric sensor was fabricated using this composite film, and the mechanical properties and electrical impedance behavior of the sensor were investigated and analyzed.A multi-channel foot pressure collection and classification system was established based on this sensor, and a Parkinson’s disease machine learning model for multi-channel foot pressure collection was investigated. Various machine learning models for Parkinson’s disease were compared, and a fine K Nearest Neighbors（KNN） Parkinson’s disease classification model with an accuracy of 97.1% was proposed. This offers a novel solution for Parkinson’s disease diagnosis and holds significant reference value.