An IoT-assisted Alzheimer’s disease patient monitoring system using adaptive deep learning models with recommendation about patient abnormality

Abstract Internet of Things (IoT) has revolutionized healthcare by enabling continuous and real-time monitoring of patients, particularly for chronic diseases like Alzheimer’s disease (AD). IoT-assisted monitoring systems leverage various technologies, such as sensors, wearable devices, and wireless communication, to monitor AD patients remotely. By integrating sensors such as electroencephalogram (EEG) monitors, Global Positioning System (GPS) trackers, and motion sensors, IoT systems can collect real-time data on the patient’s behavior, health status, and location. This data is then transmitted to caregivers or medical professionals through cloud platforms, allowing them to monitor the patient’s condition from a distance, assess cognitive functioning, and intervene when necessary. In this paper, an efficient AD patient monitoring system is introduced with deep learning. At first, the required EEG signals are gathered from benchmark resources, and then, wave and spectral features are extracted. The extracted features are then subjected to the Adaptive Dilated Gated Recurrent Unit (ADiGRU) for detecting Alzheimer patients. Here, the detection process is improved by optimizing the GRU’s parameters using the Refined Alpha Value-based Dark Forest Algorithm (RAV-DFA). If the patient is detected with AD, then the patient is monitored by CCTV. Subsequently, from the surveillance monitored videos, the patient’s abnormality is predicted by 3D Adaptive Residual Attention DenseNet (3DARADNet) and its parameters are tuned by RAV-DFA. If the abnormality is detected from the videos, then the recommendations are given to the concerned persons using Adaptive Generative Long Short-Term Memory (AGen-LSTM). For improving the performance of this phase also, the RAV-DFA is used, where the parameters of Gen-LSTM are fine-tuned. Thus, an effective AD patient monitoring system is implemented, and finally, the performance of this developed system is determined and contrasted with other related models for ensuring its supremacy.