Mark from Boston is building a sensor based application that talks HL7 (open standards) to all systems so that an air ambulance company (Boston MedFlight) can gather data in advance of hospital admission. Thus they need to build a sensor gateway. All the filtering (how many times you get information per minute from the sensor) can happen in the gateway or even in the application (so doesnt need to be in the sensor). Data mining can be done later. They’ve also built some of their own monitors and have allowed Welch-Allen and others to connect.
On top of the sensors they need an application, and they are building a GUI which shows the body etc, and will probably be run on a tablet PC to be used in the helicopter. But of course this data requires context (e.g. depending on the patient state…what if the patient is pregnant, a child, on certain meds?) and so they are building a rule processing system that prompts for the rules (that Med Flight mostly already has). Eventually they want this data to be mined and so they want to be able to collect it….again all open source standards. So their sensors are running TinyOS (sensors open source operating system). There’s a GPS in it too, so they can track location. They also want applications to exchange data — SOAP (better but incorrectly known as “web”services) in order to get these distributed pieces to exchange data. Here’s a chart of their infrastructure.
Figure 1: iRevive interface between field, 10-blade server