Episodic sampling: Towards energy-efficient patient monitoring with wearable sensors

Abstract

Energy efficiency presents a critical design challenge in wireless, wearable sensor technology, mainly because of the associated diagnostic objectives required in each monitoring application. In order to maximize the operating lifetime during real-life monitoring and maintain sufficient classification accuracy, the wearable sensors require hardware support that allows dynamic power control on the sensors and wireless interfaces as well as monitoring algorithms to control these components intelligently. This paper introduces a context-aware sensing technique known as episodic sampling – a method of performing context classification only at specific time instances. Based on Additive-Increase/Multiplicative-Decrease (AIMD), episodic sampling demonstrates an energy reduction of 85 percent with a loss of only 5 percent in classification accuracy in our experiment.

Scalable ECG hardware and algorithms for extended runtime of wearable sensors

“…Between this intervals the MCU is set to a deeper sleep mode (MCU Energy Mode 3) and the ECG frontend is powered down to save energy. This functionality has been implemented to support episodic sampling introduced by Au et al in [25]. But for this work two cases at fixed intervals has been tested: 1) A duty cycle of 2:15 (2 minutes scan each 15 minutes) and 2) a daily health check (10 minute scan once a day).…”

Context-aware energy-efficient wireless sensor architecture for body activity recognition

“…Among them we can cite sensor set selection [3]- [5], deactivation of power-hungry sensors [2], [6], [7], adaptive sampling rate [8], [9], communications reduction [10], duty-cycling [11] and resource sharing [12], [13]. However, these works are usually interested in the energy consumption of the sensors but never consider that the base station can also be energy-constrained.…”

Energy-Efficient Context Classification With Dynamic Sensor Control

“…The proposed solution leverages context awareness by modeling the dependencies between successive contexts: at a particular point in time, some contexts are stochastically more likely than others [20]. The main intuition to energy-efficient sensing is to activate sensors episodically (e.g., [28]). This paper proposes a decision-theoretic approach to select sensor resources for accurate context classification, while keeping the average energy consumption below some budget .…”