ABSTRACT
Recently, smartwatches, fitness trackers, and health monitoring devices have become popular. These gadgets collect and analyse user activity, health, and environmental data using various sensors. Artificial neural network (ANN) models process and interpret wearable sensor data to enable intelligent decision-making and personalized experiences. This study reviews clever wearable ANN models. We review the literature and divide ANN applications into three primary categories: activity recognition and prediction, physiological parameter estimation and monitoring, and user behaviour analysis. ANN models have been used to identify and predict human activities using accelerometer and gyroscope data. These models improve users’ well-being and performance by tracking activities, contextualizing, and providing personalized feedback. ANN models are essential for physiological parameter estimates and monitoring in intelligent wearables. ANN models can provide real-time health status monitoring, abnormality diagnosis, and personalized health recommendations by analysing sensor signals from heart rate monitors, electrocardiograms, and blood pressure sensors. User behaviour analysis, including sleep tracking, stress detection, and dietary assessment, uses ANN models. These algorithms can identify user behaviours, hazards, and healthy habits by learning sensor data patterns and correlations. Intelligent wearables use feedforward, recurrent, and convolutional neural networks. We also discuss ANN model optimization, privacy, and interpretability in wearables.
