Improving the Efficacy of Continuous Monitoring of Patients with Parkinson's Disease Through the Use of an Intelligent Computing Architecture Within the Medical Internet of Things

Abstract

Telemonitoring has grown in popularity these days, particularly to assist patients with serious illnesses such as Parkinson's disease (PD). This study relies on the Daphnet dataset, which was trained and utilized to track five patients' whereabouts for an extensive dataset. Fuzzy logic was used in conjunction with a linear and Mobius map as part of the fog computing freezing of gait (FoG) detection system to provide a multi-level output (MLFM-map) that takes advantage of various spatial resolutions in motion data processing. Two improved Salp Swarm Algorithms with Fuzzy Logic (Fuzzy-ISSA) and transient search optimization algorithm (TSO) have been employed for the methodical training of this tool at the level of fog computing. When compared to other machine learning techniques, this one trains the FOG detection system quite quickly. In order to maximize performance while lowering computational complexity and testing time, the model architecture and parameters have been carefully considered. On average, the suggested method detected over 90% of FoG occurrences with extremely low latency in the original (test) dataset, demonstrating good to exceptional classification performance. In addition, the algorithm demonstrated a specificity of over 90% when tested on the test set. Various algorithms are used to train this design.