Are armband sEMG devices dense enough for long-term use?—Sensor placement shifts cause significant reduction in recognition accuracy. (July 2020)

• Record Type: Journal Article
• Authors: Kanoga, Suguru; Kanemura, Atsunori; Asoh, Hideki
• Is Part Of: Biomedical signal processing and control. Volume 60(2020)
• Publication Date: 2020-07
• Subjects: Myoelectric control system -- Electromyogram (EMG) -- Wearable sensor -- Long-term use -- Motion recognition; Signal processing -- Periodicals; Biomedical engineering -- Periodicals; Signal Processing, Computer-Assisted -- Periodicals; Image Processing, Computer-Assisted -- Periodicals; Biomedical Engineering -- Periodicals; 610.28
• DOI: 10.1016/j.bspc.2020.101981
• Languages: English
• ISSNs: 1746-8094

Abstract

Highlights: Armband myoelectric control systems were developed. Armband sEMG device was used for capturing 22 types of forearm motions. Performances of sEMG recognition algorithms were evaluated. Armband sEMG devices are dense enough for short-term use but not apt for long-term use regarding the conventional recognition algorithm. Abstract: Myoelectric control systems (MCSs), which recognize motions through surface electromyograms (sEMGs), present potential applicability for clinical, recreational, and motion-assisting purposes. To increase the adoption of armband device-based MCSs, the performance of motion recognition algorithms should be determined over long periods and sensor placement shifts. We prepared an sEMG dataset to assess motion recognition algorithms for practical use over long periods with varying sensor placement. The dataset comprises 30 recording sessions over 40–42 days, in which sensors were placed at three different placements. We used an armband eight-channel sEMG device for capturing 22 types of forearm motions from five healthy male subjects. To consider only motion periods to learn classifiers, we extracted relevant 1.5-s segments via multiscale sample entropy. We evaluated the dataset on a conventional motion recognition algorithm, finding robust intraday performance but significantly deteriorated inter-day performance under varying sensor placement. Hence, the armband sEMG device is dense enough for short-term use but not apt for long-term use regarding the conventional recognition algorithm. Adaptation techniques are required for developing armband device-based MCSs for long-term use. The dataset and sample codes from this study are publicly available at GitHub .