AI system cuts energy use in aging-friendly homes while tracking daily activities
Researchers have developed an artificial intelligence framework that monitors elderly residents' daily routines through smart home sensors and automatically optimizes heating, water, and waste systems to reduce carbon emissions. As the global population over 60 doubles to 2.1 billion by 2050, the technology offers cities a scalable way to improve senior care while meeting sustainability targets.
Originaltitel: Smart home optimisation for sustainable cities: The AiCareASDL framework for resource management and carbon footprint reduction
<p>The growing elderly population whose age is above 60 years is approximately 1.1 billion people globally, as of the data taken for 2022. This constitutes 13.9% of the total population. The elderly population is predicted to double over the coming years. and will reach around 2.1 billion by 2050. As age advances, there are many limitations faced by the elderly to manage daily activities; this will indirectly influence environmental sustainability and carbon emissions. To enhance the general quality of life for urban residents, the proposed work “Artificial Intelligence Care Activities for Sustainable Daily Living” (AiCareASDL) deploys Gaussian fuzzy temporal extraction from IoT sensors to study activities of daily living. This data is used to derive patterns of daily activities using Takagi-Sugeno-Kang fuzzy classifier. The fuzzy pipeline is used to classify sleep duration index, wellness index, energy consumption, water, and waste management. The proposed research interprets the sensor matrix to analyze how inherent uncertainties and variabilities arise within the system. The AiCareASDL architecture is implemented across three primary levels: the first deploys the sensing IoT units; the second processes data using a fuzzy algorithm that incorporates fitness computation and crossover operators; and the final level recognizes activity patterns once predefined criteria are met. Sustainable living is achieved by calculating resource usage for water and waste management. AiCareASDL implements the statistical analysis of water management used in cooking and cleaning activities at optimal and suboptimal sensor levels. Waste management for empty, medium, and full of dustbin. Deep learning is implemented in addition, as it can identify latent patterns in raw sensor data that have been missed by rule-based fuzzy systems. The data from IoT sensors is processed into deep learning models like VGG19, ResNet50, Inception V3, MobileNet and DenseNet 121. The performance metrics of accuracy, upper bound, lower bound, and sensitivity are calculated and conclude that DL models performed better than the fuzzy classifiers with 95% accuracy. The approach’s concrete benefits are illustrated by case studies that show how it enhances urban residents’ quality of life, lessens its impact on the environment by deriving the sustainable parameter for carbon footprint. By encouraging the harmonious coexistence of ecological preservation and urban growth, AiCareASDL makes intelligent and sustainable cities a reality.</p>