Cloud-Connected Bracelet for Continuous Monitoring of Parkinson’s Disease Patients: Integrating Advanced Wearable Technologies and Machine Learning

Asma Channa, Giuseppe Ruggeri, Rares-Cristian Ifrim, Nadia Mammone, Antonio Iera, Nirvana Popescu

Parkinson’s disease (PD) is one of the most unremitting and dynamic neurodegenerative human diseases. Various wearable IoT devices have emerged for detecting, diagnosing, and quantifying PD, predominantly utilizing inertial sensors and computational algorithms. However, their proliferation poses novel challenges concerning security, privacy, connectivity, and power optimization. Clinically, continuous monitoring of patients’ motor function is imperative for optimizing Levodopa (L-dopa) dosage while mitigating adverse effects and motor activity decline. Tracking motor function alterations between visits is challenging, risking erroneous clinical decisions. Thus, there is a pressing need to furnish medical professionals with an ecosystem facilitating comprehensive Parkinson’s stage evaluation and disease progression monitoring, particularly regarding tremor and bradykinesia. This study endeavors to establish a holistic ecosystem centered around an energy-efficient Wi-Fi-enabled wearable bracelet dubbed A-WEAR. A-WEAR functions as a data collection conduit for Parkinson’s-related motion data, securely transmitting them to the Cloud for storage, processing, and severity estimation via bespoke learning algorithms. The experimental results demonstrate the resilience and effectiveness of the suggested technique, with 86.4% accuracy for bradykinesia and 90.9% accuracy for tremor estimation, along with good sensitivity and specificity for each scoring class. The recommended approach will support the timely determination of the severity of PD and ongoing patient activity monitoring. The system helps medical practitioners in decision making when initially assessing patients with PD and reviewing their progress and the effects of any treatment.