DESIGN
EW: MEDICAL
models at the Edge, MIoT devices can drive localised decision-making without relying on communication with centralised infrastructure. Furthermore, with Edge processing, communication can be selective and intelligent, transmitting critical information only when it is required, helping to reduce network traffic. Applications and benefits of Edge computing in MIoT While the potential for MIoT devices equipped with Edge AI and ML models is extensive, some key applications already show promising developments to help increase the effectiveness of healthcare institutions. Remote patient monitoring The number of remote patient-monitoring (RPM) solutions globally will increase from 75 million in 2023 to 115.5 million by 2027, equating to 1.4% of the total population, according to Juniper Research. Edge-enabled RPM devices can continuously monitor vital signs like heart rate, oxygen saturation, or glucose levels while adding intelligent insight and providing real-time alerts if a patient’s condition deteriorates. This reduces the need for constant human oversight and improves patient outcomes through timely interventions. Real-time diagnostics Edge processing in MIoT can help drive real-time diagnostics by collecting comprehensive patient information and comparing it against pre-existing datasets. For example, a wearable or stationary MIoT device monitoring a patient’s elec- trocardiogram (ECG) can instantly identify arrhythmias without needing to send the data to a central server. This enables efficient diagnostics without the need for a healthcare professional and reduces delays, which is especially critical in time-sensitive emergencies.
Advanced Edge processing will also drive the concept of adding intelligence to existing medical equipment to extend its functionality. Devices such as smart insulin pumps (Figure 1) or automated defibrillators benefit from Edge computing by perceiving and, therefore, instantly responding to patient needs. These devices can adapt their functionality by adjusting dosages or shock timings based on real-time data analysis, enhancing patient care and safety. Adding intelligence to a device can also enhance preventative maintenance, as Edge intelligence enables MIoT devices to predict service needs and prevent costly downtime. Through continuous monitoring, Edge ML models can spot changes in medical equipment that might indicate the need for maintenance, This, in turn, helps to reduce unexpected equipment downtime, maintenance costs, and healthcare risks. Large language models At present, deploying full Generative AI models based on large language models (LLMs) to low-powered hardware is not feasible due to the significant computational resources they require. However, ongoing research in model compression and optimisation techniques may enable smaller, task-specific models to be deployed on such devices in the future. These advancements could lead to the development of smarter medical assistant solutions. For example, MIoT-enabled devices at a patient's bedside could use AI models to process spoken requests without a Cloud connection, before relaying them to healthcare teams. This would streamline response times by providing healthcare personnel with critical patient information before attending to them, enhancing efficiency and patient care. Essential electronic components for MIoT devices Selecting the right components is
Intelligent medical devices
30 ELECTRONICSPECIFIER.COM
Powered by FlippingBook