Internet of Things - Remote Sensing in  Healthcare

The project focuses on improving healthcare in developing countries using low power, economic electronics such as the Raspberry Pi (Rpi).

As a student originally from Kenya, I appreciate the level of healthcare available through organisations such as the NHS. In the 21st Century, they make use of extensive networks to gather and share patient monitored data, which is used to perform analysis and provide better healthcare services.

Using devices such as the Raspberry Pi, I aim to help create a system where data can be transported between remote locations, where internet connection is limited - a physical network of sorts, if I may.

The system utilises 3 key components:
1. Nodes that gather data in the field. These are Rpi's used in small clinics in rural towns. Each clinic can have a node, or on a larger scale, each hospital room/wing could have a node. The node is connected to sensors such as heart monitors, blood pressure monitors, etc. I only had access to a webcam so I have modelled this as a sensor to monitor patients movement.

2. Hubs that collect all the data from different nodes and upload it online. The hub is located in a major city with internet access.

3. Microsoft Azure Cloud Services - This is where the data is uploaded and processed. An upcoming technology in the Internet of Things, the Azure Service Bus is a great messaging service to transport messages and large data, as well as store logs in the form of tables.

The data from the Service Bus can then be accessed from anywhere in the world. The purpose of doing this is so that the patient data/information can be processed in more developed countries and feedback can be sent to the individual nodes through the hubs. On a more local scale, some of the sensor data could be used to trigger messages on the service bus such as health alerts that could be sent to nurses to act upon.

There are a range of scenarios of data transportation from nodes to hub, depending on the availability of transport resources:

1. Availability of Cars/Road Network: In this case, a person could be assigned to drive through the nodes to gather the data and drop it off at the hubs. Here a simple USB drive/external SSD could be used.

2. No road transport available: Here, a UAV or something similar could be used to transfer the data between nodes and hubs. Research that I carried out during the Summer of 2014 at the University of Southampton investigated this option. The UAV would have a wireless SD card (see BoM) that would automatically receive the data from the nodes, then fly to the hub, which would automatically download the data from the UAV. (Note that the UAV could be recharged using wireless induction charging as it waited to receive data from nodes.)

At the moment, the project is a small scale model that works with both scenarios, but the hope is to get the platform tested and receive some funding to take it to an industrial and applicable level.

The innovation in this project lies in the automation involved, where no human interaction is required, other than of course plugging in the USB, or controlling the UAV. The other main factor is the integrated use of Microsoft's Azure Service Bus, which is a great piece of technology enabling the Internet of Things on a large scale. With the new Raspberry Pi 2 enabling the use of Windows 10 on it, the platform could be further integrated with this technology.

An adaptation of this project for wildlife conservation (written by myself) is published on 

http://udoo.hackster.io/umangrajdev/internet-of-things-remote-sensing

with more details on 

http://sjj.azurewebsites.net/?page_id=2171

If you have any queries, would like to discuss the project, or have any opportunities for funding, please do contact me! Thanks for reading, and if you like the project, do add some Respect using the links on the side.

*Note the build logs below are in reverse chronological order. I have them written out as tutorials, and it would be best to read from the first log onward to the last.* 

*A lot of the text is taken from my blog on remote sensing for wildlife, since the set up is very similar.*