I-Corps: A Smart Technology for Child Safety

点击翻译按钮获取中文摘要

The broader impact/commercial potential of this I-Corps project is focused on a smart technology that will help smart tethering between elements of social dyads such as parents/infants, caregivers/ patients, or even owners/pets. Given a complex set of scenarios in which such dyads operate, in some subset of scenarios alerts must be generated, while in other complex scenarios no alerts should be generated. This project employs machine learning techniques to propose a robust technology that minimizes false alarm rate. The system will automatically learn about the environment (unsupervised learning) to minimize the false alarm rate. The smart tethering system consists of hardware and smart software algorithms. Possible applications include child monitoring, patient monitoring and possibly other safety-motivated interventions.This I-Corps project will further develop a smart tethering technology that integrates sensors, computing platforms, and artificial intelligence. The proposed technology implements a combination of genetic algorithm (GA) and deep learning approaches to model the environment to minimize the response error. Initially the system is trained with training data (GA coupled with deep learning) and then deployed in user's environment. After deployment, the system smartly learns complex scenarios and customizes to the user's environment based on unsupervised learning approach. The project builds upon prior work in hardware development for smart adaptive systems. In general, for smart solutions, researchers mostly have employed supervised learning (training based) due to its simplicity. If this research is successful, it will advance the state-of-the-art in the unsupervised learning approach as well as GA based smart systems.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.