My Diabetes Smart Home

Index page for my research projects.     Last updated Nov 2021

Nightscout xDrip+

Download latest APK  

Nightly snapshots
Source repository
Personal research version of xDrip


Adaptive and Predictive Alerts
Adaptive and Predictive Alerts
Treatment Simulations in real-time
Treatment Simulations in real-time
Treatment Input by Voice or Keypad
Treatment Input by Voice or Keypad
and via Android Wear SmartWatch
and via Android Wear SmartWatch
Undo / Redo History + Annotations
Undo / Redo History + Annotations
Time-block Profile Editor
Time-block Profile Editor
Live Sync with Follower Handsets
Live Sync with Follower Handsets
Customizable + many Data Sources
Customizable + many Data Sources
Full Nightscout Sync via NS-Client
Full Nightscout Sync via NS-Client
Choose your own colors
Choose your own colors
Built-in Translation Editor
Built-in Translation Editor
Multiple Language Support
Multiple Language Support
Tasker Scripting Plugin
Tasker Scripting Plugin
All Parakeet Features Integrated
All Parakeet Features Integrated

The Parakeet

The Parakeet is a portable home-built device which receives wireless signals from a commercial continuous G4 glucose sensor worn on the body. It transmits these over the phone network to a private or cloud internet server. It does not work with the later G5 and G6 transmitters.

The Parakeet is primarily designed to allow parents and carers to be able to monitor the blood sugar of a diabetic child even if they are a long distance away, for example, carried in the pocket of a school bag. The real-time blood sugar information would be available on the parent’s mobile phone — potentially many miles away.

Whole house receiver coverage

Using multiple Raspberry Pi silent mini-computers combined with Wixel receivers we can achieve blanket coverage of the whole house. It is no longer necessary to carry a CGM receiver or remain in range of one when at home. The signals can be collated and interpreted by these networked machines.

USB Wixels can use the Parakeet codebase, simply set use_gsm = 0 when compiling.

Automation Systems

Sophisticated forward prediction and anomaly detection systems, linked with internet connectivity, make it possible to provide alerts and computer based analysis intelligently. This can minimize “alarm fatigue” and reduce the mental burden of constantly tracking blood sugars.

Just about anything is possible once we have open interoperation of the glucose sensor data.

Currently there is active testing of systems which produce alerts based on sensor heuristics and probability tuned predictions of what is likely to be occurring at a given moment. Machine learning systems which can interpolate a lifetime of blood glucose data, detect anomalies and escalate alerts as needed.

These may range from a simple ambient lighting change indicating a blood sugar prediction right up to an emergency call out to friends and family if the diabetic becomes unresponsive.


Contact Information

To contact me, with questions or for commercial work you can reach me via email jamorham /at/ gmail.com