Wireless Body Sensor Network for Continuous Cuff-less Blood Pressure Monitoring

Abstract

This paper describes an unobtrusive IEEE 802.15.4-based wireless Body Sensor Network (BSN) that enables continuous cuff-less blood pressure monitoring, opening up new perspectives for hypertension diagnosis and treatment, cardio-vascular event detection, and stress monitoring. We estimate the arterial blood pressure based on the Pulse Arrival Time (PAT), which is measured with a waist electrocardiogram (ECG) and an ear photo-plethysmogram (PPG). The PAT measurement requires the synchronization of the wireless sensors' clocks. This is achieved with the Flooding Time Synchronization Protocol (FTSP). The evaluation of both the implemented time synchronization and the robustness of the wireless data transmission yielded promising results. Future work will include the study of packet collisions in synchronized IEEE 802.15.4 networks, the enhancement of the BSN with additional context-providing sensors (e.g. accelerometer, magnetometer, thermistor), and the integration of the blood pressure estimation algorithm in a wireless sensor unit.

A body-fixed-sensor-based analysis of power during sit-to-stand movements

“…For node synchronization the flood time synchronization protocol was used: one node in the network is the time master and is updating periodically the time for every other node (cf. [11]). Synchronous data of all sensor nodes were obtained at a sampling rate of 50 Hz.…”

Experiences with Pulse Arrival Time as Surrogate for Systolic Blood Pressure

“…We have realized a couple of measurement systems tailored for wearable continuous monitoring scenarios. One prototype realized simultaneous detection of an ECG and a peripheral Photoplethysmogram (PPG) [5] including activity /posture inference with a body sensor network (BSN). The BSN consists of two wireless sensors, which transmit data to a central unit (PDA) on the basis of the low-power short-range standard IEEE 802.15.4.…”

PDhms: Pulse Diagnosis via Wearable Healthcare Sensor Network

“…For example, feature extraction is referred, but not discussed in Ref. [11] and [12], because that these work makes the assumption that sensed data is rarely mixed with noise or disturbance. Although lots of existing work collects blood pulse data for a healthcare-monitoring purpose, the data is often analyzed semi-automatically [8], and pulse diagnosis is conducted manually [11].…”