Designing Multi-sensory Personal Aids in the IoT Age

点击翻译按钮获取中文摘要

This program targets quality-driven design of Internet of Things (IoT), with emphasis on health/wellness applications built on multi-sensory wireless systems. Design and validation will be explored for secure and safe systems including dependable, inexpensive and usable Machine Intelligence (MI) incorporation on the network edge. In long-term, our advances should enable the personalized medicine systems, where major physiological state variables are tracked to personalize medicaments, exercise and nutrition dosing, leading to precise control that reduces widespread adverse effects, especially for chronic conditions. Short-term goals include the quality-driven design of dependable systems, the development of health testbeds and the trust/certification compliance. Systems for chronic health conditions (diabetes, respiration) and aging will be tuned to improve or obviate the (separately undertaken) human trials. The increasing use of MI creates problems with dependability and with being explainable (trusted), such that they can be safely added to health and wellness devices. Accelerators will be designed for distributed, privacy- preserving human activity recognition (HAR) and other MI, as well as harnessing the carefully designed new capabilities of the devices embedded in the edge of the network. To achieve diverse quality goals, we will combine our property monitors with domain-specific tools. Ensuring proper data handling by diverse stakeholders is currently costly, as common IoT platforms become security targets, which especially endangers privacy in health and wellness applications, towards which we will undertake several threads of research in secure design, trust and safety. Besides the assertion-based quality assurance, monitoring and verification techniques, we will explore the repudiation-free blockchain for feasible trust mechanisms in networks of IoT devices used in healthcare and wellness, including notably Body Sensor Networks (BSNs). Scalable blockchain technologies will be developed for the realization of electronic health records addressing the challenges of providing trusted prescriptions, medicaments and sensor data in the exchange between medical practitioners, pharmacies, insurance companies and patients, allowing the health/wellness regimen and, ultimately, precision (personalized) medicine. Research focus on the usability and continuous service quality improvements of IoT for healthcare will be hence combined with the security and privacy concerns. We will explore the use of new hybrid system optimization and verification methods, as well as the emerging computing fields, such as the imprecise, in-memory, reversible or quantum computing. Future-proofing of the research will include the quantum computing-resistant security and scalable reversible circuit synthesis, with benefits ranging from energy-efficiency to intelligence, safety and privacy preservation. We will train 20-30 HQPs in high demand for industry and academy.