Executive Summary : | The demand for remote diagnosis of patients is increasing day by day. Recently, patient diagnosis integrating the Internet of things (IoT) with cloud-based technologies has created wide attention in the research community. However, a unified IoT-cloud-based framework that supports hand-held point-of-care devices (POCDs), wearable devices, and existing devices (e.g., CT scan) and features a pandemic-compliant infrastructure is lacking. In this project, three categories of the preventive healthcare system will be integrated with the cloud and IoT framework: a) self-designed and developed hand-held/wearable devices, b) existing devices empowered with intelligence, and c) pandemic-compliant infrastructure. The medically underserved areas in the low- and middle-income countries (LMICs) are vast and infrequently engaged with the health system. In order to remotely give patients and healthcare professionals the detection and diagnostic capabilities, realistic POCDs are required. The proposed study seeks to build a field-type hand-held POCD that can record several bio-signals (such as heartbeat, ECG, EMG, and EEG) and communicate the pertinent data to distant hospitals (or medical practitioners). Moreover, an intelligent algorithm will be added to the POCD to assess and characterize the bio-signals and offer immediate diagnostic measures. Although POCDs promise to provide a viable solution, such technologies have not succeeded in penetrating the masses. Suspicion of the machine-based outcome and machine-related dread are the main barriers to accepting such devices. In order to address this issue, we suggest implementing a POCD with medical assistance, i.e., doctor-assisted POCD (DAPOCD). Using an application program, medical professionals can access the patient's information and offer prognostic measures (using a Smartphone-based App or Web). Further, the methodology will be extended to wearable healthcare devices. Additionally, along with integrating DAPOCDs and wearable devices to design a smart healthcare ecosystem, the existing devices (e.g., high-resolution computer tomography (HRCT)) in healthcare will be equipped with intelligence. Furthermore, the framework will be resilient to face any pandemic in the near future by maintaining pandemic protocols (proper type and proper wearing face mask, social distancing, etc.) and detecting physiological parameters (fever, voice, oxygen level, heart rate, etc.) of the infected individuals. Typically DL algorithms used to classify bio-signals cannot be integrated into memory-constrained hardware such as microcontrollers which restrict their use in hand-held field-type devices. More recently, with the dawn of TinyML (Tiny Machine Learning (ML)), the DL algorithms can be compressed significantly so that they can fit in the microcontroller without losing their classification capability. Incorporating TinyML will be actively sought in the proposed research to design and develop an intelligent healthcare framework. |
Co-PI: | Dr. Kandarpa Kumar Sarma, Gauhati University, Assam-781014, Dr. Anjan Kumar Talukdar, Gauhati University, Assam-781014, Dr. Jyoti Prakash Medhi, Gauhati University, Assam-781014 |