A Decision Analytics Approach to Intelligent Smart Home for Sustainable Healthy Living

Zhou Ruohua

Abstract

Problems pertaining to the aging global population are becoming increasingly serious, with the United Nations (2018) predicting that they will be among the most important social issues of the century. For instance, Singapore has experienced a large growth in the proportion of its elderly. One problem resulting from the aging population is that current public health systems will be unable to fulfill the needs of the elderly in the near future. A number of studies have addressed home care solutions as a complement to hospital care (Bennett, Rokas & Chen, 2017). However, these studies have either focused primarily on motion analysis and implementation or have provided a simplified version of a rule-based decision-making model for comprehensive disease analysis.

This study proposes a comprehensive decision-making model to handle both time-critical (reactive) and computationally-intensive (predictive) decision-making in healthcare for the elderly. The proposed intelligent smart home includes four interconnected systems: an intelligent-reasoning and decision-making system, a cloud system, an external system, and external participants. This paper focuses on developing a decision analyzer in the intelligent reasoning and decision-making system. This novel method extends the existing smart home decision analyzer to two intelligent reasoners with the help of influence diagrams, which can handle the decision under uncertainty and are easily combined into larger influence diagrams. A reactive decision analysis unit is responsible for time-critical decision analysis, while a predictive decision analysis unit predicts long-term risks given an individual's current health condition.

With this setup, a smart home intelligence system can quickly respond to conditions such as heart attacks and hypertensive disease and process a comprehensive risk analysis when necessary. This enables at-home health monitoring, freeing a large amount of public resources. Early reaction to and detection of diseases has the potential to greatly increase the survival rate and life expectancy of the patients. Furthermore, a smart home intelligence system could reduce false-negative rates by collecting and updating multiple evidences with different symptoms of the same disease. The proposed system is easily adjustable by data engineers and medical professionals based on progress in frontline research.

In summary, the proposed system is a step towards a truly intelligent smart home system. It integrates achievements in separate fields, including medical research, motion analysis, and cloud-based smart home environments (CoSHE), in an effort to achieve optimal effectiveness.