Tremor stabilization improvement using anti-tremor band: a machine learning–based technique

Introduction: Tremors, the most prevalent movement disease, restrict sufferers from going about their everyday lives and participating in physical activities, ultimately lowering their quality of life. Due to their pathophysiology, developing effective pharmacological therapies for tremors is difficult, which could be more efficient in regulating tremors. Thus, various treatments are needed to address this developing, age-related illness. With surgical methods like deep brain stimulation, tremors can be controlled. However, its utilization is only partially used due to high prices, patient and doctor choices, and considered significant dangers. Medical devices are well-positioned to bridge the gap between surgical procedures, pharmaceutical treatments, and lifestyle changes to achieve effective and safe tremor control. Materials and methods: The wearable IoT-based anti-tremor band is proposed to address this issue, fulfilling the patient’s everyday needs. Wearable anti-tremor bands function by internally producing pressures that neutralize or reduce the intensity of the user’s tremor. Targeted by the wrist-worn device are the radial and median nerves. An integrated accelerometer sensor detects motion or handshaking and operates the motor appropriately. By analyzing accelerometer data and sending real-time control signals to the vibration motors, the band uses machine learning techniques to stabilize vibrations. Results: The suggested system shows efficacy in lowering tremor severity and improving control through thorough experimentation and analysis. According to the findings, the created anti-tremor band has great promise as a non-invasive and reasonably priced treatment for those with tremor-related diseases. Conclusion: To locate tremors, the method uses accelerometer data collected in the field. A unique tremor label and a collection of acceleration signal segments correspond to each individual. Comprehensive testing on a patient and non-patient dataset confirms that these occurrences can be identified with the suggested method. © 2023, The Author(s), under exclusive licence to The Brazilian Society of Biomedical Engineering.