Key result
The SubRISC+ processor with a memory-conscious dynamic time warping algorithm achieved significantly lower circuit area and power consumption compared to state-of-the-art embedded processors.
Why the study?
Continuous real-time disease diagnosis using portable IoT eHealth devices is limited by constant power demands and heavy resource constraints on battery-powered systems.
A memory-conscious dynamic time warping algorithm implemented on a low-power embedded processor provides a power-efficient solution for continuous eHealth monitoring devices.
May support efficient eHealth hardware designs; leaves open clinical validation and outcome trials.
The rapid development of Internet of Things (IoT) has opened new opportunities for healthcare systems through so-called eHealth systems. With the help of monitoring using portable IoT devices with biomedical sensors, disease diagnoses can be conducted in real time. However, there is a challenge in that monitoring is an always-on activity that requires constant power supply and IoT devices are battery-powered and face heavy resource constraints. This work addresses a realistic implementation of a low-power eHealth device using both hardware and software approaches. We realize various lightweight eHealth applications (particularly monitoring applications) using a memory-conscious dynamic time warping (DTW) algorithm to be deployed on a small and low-power embedded processor. Actual prototypes of the processor are currently being fabricated. Our evaluation showcases the effectiveness of our work compared with other state-of-the-art embedded processors in terms of circuit area (fabrication cost) and power efficiency. We also demonstrated the scalability of the software implementation by varying the amount of data used in DTW for various eHealth monitoring applications.
No takes yet. Share an insight, caveat, or question.
Yang et al. (2020) studied this question. SubRISC+ embedded processor with memory-conscious dynamic time warping (DTW) algorithm vs. ARM Cortex-M0, RISC-V zero-riscy, and RISC-V micro-riscy processors was evaluated on Circuit area (fabrication cost) and power consumption. The SubRISC+ processor with a memory-conscious dynamic time warping algorithm achieved significantly lower circuit area and power consumption compared to state-of-the-art embedded processors.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: