Communicating Mobile Biometric Sensors

Objectives

In project CoMoBioS we investigate the possibility of using off-the-shelf wireless sensors to monitor the health of highly mobile people in outdoor conditions. To achieve this goal we focus on a scenario we consider as a most challenging one: monitoring the cardiac activity of runners during a marathon race. This particular scenario was selected because runners must cover a long distance during a marathon, and that distance clearly exceeds the limited radio range of the low-power radio transceivers available on most current sensor platforms. Besides, since runners in a marathon all follow exactly the same route, a number of base stations can be deployed along that route.

A base station (BS) is typically a unit that features a low-power radio interface, and at least one wired or wireless interface for long-distance transmissions (typically a broadband access to the Internet). The first radio interface is used to receive data from the sensors carried by marathon runners, and the second one is used to forward these data to a remote site (for example the closest medical aid station, or a physician’s desktop, laptop, or smartphone). Data received from the sensors can be processed locally on the BS before being forwarded to the monitoring site, although that is not a requirement.

Deploying hundreds of base stations in order to cover a marathon route is hardly an option, for organizational and financial reasons. The approach we propose is based on the assumption that only a sparse coverage of the route needs to be ensured, using a reasonable number of base stations. A disruption-tolerant solution for data gathering must therefore be implemented, using the store and forward principle, which is the foundation of Disruption-Tolerant Networking: a mobile node that is temporarily disconnected from the network can store data (or messages) in a local cache, carry these data for a while, and forward them later when circumstances permit.

In our scenario, the ECG sensor carried by a runner captures data continuously and stores these data locally. Whenever the runner passes by a BS, a transient radio contact occurs between the sensor and that BS. This contact is exploited by the sensor to upload data to the BS, which in turn can relay these data to the monitoring center. With this approach, the marathon route can be covered satisfactoritly with about 30 base stations. For an average runner, this means that “fresh” data about this runner’s cardiac activity can be received at least every 5 to 10 minutes by the remote monitoring center.

In this project we use Shimmer™ platforms with ECG expansion modules for data acquisition. Besides, several technologies have been considered in order to support the episodic transmission of ECG data between sensors worn by runners and roadside base stations.

Results

The IEEE 802.15.4 (ZigBee) technology was originally considered for sensor-to-roadside base stations, but field experiments conducted during an intra-campus sports event revealed that 802.15.4 transmissions can hardly meet the requirements of the marathon scenario. Details about this experiment can be found in this paper.

Since ZigBee transmissions proved ineffective in our marathon scenario, we investigated an alternative approach that involves using Android smartphones as relays between ECG sensors and roadside base stations. With this architecture, IEEE 802.15.1 (Bluetooth) transmissions are used on the sensor-to-smartphone segment, and IEEE 802.11 (Wi-Fi) transmissions on the smartphone-to-base-station segment. A second round of field experiments has confirmed that this approach is viable, and that it allows to transmit ECG episodically with no data loss. Details about this experiment can be found in this paper.

Perspectives: from marathon runners to ambulatory patients

The marathon scenario has deliberately been selected as a most-challenging test case. The underlying idea is that solutions that can meet its stringent requirements may be reused in less constrained situations.

As an example, we investigated the possibility offered by the many Wi-Fi hotspots available in our environment to collect data from biomedical sensors worn by ambulatory patients during their daily activity (patient at home, at work, shopping, practising sports, etc). Details about experiments we conducted along that line can be found in this paper.

Publications