December 2, 2019 | Written by: Dean Clarke, Rosie Lickorish, and Pablo Ortega
On the 21st of November 2019 ETS hosted PACE for a 2 day hack event. The PACE Centre is a UK-registered charity and special school based in Buckinghamshire which specialises in helping children and young people with motor disorders, like cerebral palsy, which is estimated to effect 30,000 children in the UK.
PACE transforms the lives these young people by giving them a chance to learn and unlock their full potential by offering unique schooling and therapy uniquely tailored to each childs needs, which may vary drastically due to the nature of their disability.
One of the solutions PACE use to help their students access the curriculum is specialised switches which can be placed on different areas of a persons body, depending on where a child has the best motor control, so they can easily activate it based on their ability. These switches are controlled by an interface, a grey box with 6 headphone jack style ports, that allows the switches to be somewhat reconfigured by physically changing which ports the switches are plugged into. The interface then converts the activation of the switch to a keyboard input and sends it to a receiver which can be connected to devices such as a computer, wheelchair, or lights.
While the as-is solution is already extremely powerful in enabling people with limited motor functions to interact with the world around them, significantly increasing their independence, we identified some significant limitations.
- Cost – The current switches and interface are very expensive (£50 per switch, £150 per interface) and requires many separate components to be wired together.
- Customisability – Switches are limited in customisability in both
- Hardware: you’re limited to off the shelf designs which may not suit all disabilities.
- Software: if you require something unique, i.e. a button register as being held down for 5 seconds on input, and it’s not one of the interface ports then it cannot be done.
- Design – The current equipment consists of several grey boxes wired together. These look very medical and can make the users, often children, afraid to use them.
For our solution we used a Raspberry Pi, Node-Red, a TI Sensortag, and 3D printing to design and create a prototype in 2 days.
Using 3D printing we were able to create a custom grip style switch casing to house a sensortag, by using the light sensors and accelerometers on the sensortag we were able to de
tect when the grip was being held, and when it was being shaken. These inputs were sent to the raspberry pi, via bluetooth, and injected into a node-red flow. This flow allows the user to fully customise the switch via changing things like the sensor activation thresholds, which key the input should interpreted as, the length of time to hold the key down, and more. By setting up the Pi to emulate a bluetooth keyboard we were able to send this input to a computer to control a simple program.
By using node-red in this way we are able to remove the physical wires and replace them with virtual ones, allowing for the possibility of more powerful customisation, with basic coding skills, without removing the ease of use for none coders.
By the end prototype proved there are solutions to all addressed problems:
- Cost – Raspberry Pi and Sensor tags are already significantly cheaper than the current off the shelf equipment and we can further reduce this cost via the use of custom sensors.
- Customisability – for both
- Software: using node-red we are able to fully customise the function of a button or sensor.
- Hardware: using 3d printing we can create customised cases/switches for the button which can be easily tailored to each individual.
- Design – By removing the physical wires and grey boxes the final sensors that the kids will receive look more like toys and a lot less like scary medical equipment.
The ability to connect other custom sensors via bluetooth was important, as it opens up the possibility of maintaining compatibility with the already purchased buttons via a bluetooth enabled breadboard, similar to the current interface, which would allow users who are used to the current solution have a familiar way of interacting with the new solution.
The final day ended with a proof of concept prototype that addressed all of the problems identified with the as-is solution. By working together with PACE we were also able to share our skills with node-red to help them move forward in developing the prototype into a fully realised solution.
For a future solution adding a additional UI layer over the node red flow could make more commonly used customisations more easily accessible to a non-technical user.