Exciting preliminary news in hacking the Omnipod for closed loop

Excellent idea!! I read some of the above and thought I’d mistakenly landed in the German thread!

4 Likes

Here comes Debby Downer.

So my concern is that Dexcom will do the same thing that Minimed did. In other words, they will improve the encryption of the communications protocol they use between their pod and pdm. Minimed did this soon after it was publicly disclosed at a hacker conference that it was possible to control their pump remotely, for which they undoubtedly received what they considered negative press (though it also planted the seed that eventually led to OpenAPS). Soon after that they released their next firmware upgrade - which is why you need to use an old Minimed pump that was released prior to that firmware upgrade. Luckily their old pumps have a long lifetime (several years at least), so there are still old pumps available.

This problem will be worse for the Dexcom, however, because the pod is disposable. So when they improve their encryption (I find it hard to believe they won’t?), the pods that support the known protocol will soon age out and no longer be available. Which means that this race to decrypt will need to occur again, but it will be harder each time.

1 Like

Since PDM’s have a 4 year warranty, the old pods will need to be available for 4 years from the time they start selling the new Dash PDM. That is a requirement for Insulet, they can’t wiggle out of that.

So, 1 week before the new Dash PDM is available, I am buying another one of the current PDM. That allows me 4 years of having the old system.

I suggest everyone consider the same - buying a current PDM right at release, and helping to ensure the 4 years of current encryption.

4 Likes

I’ve just checked the OmniAPS slack channel - the key comm protocol steps have been verified by multiple people, and there are already discussions about writing drivers - so, indeed excellent and exciting news! If you do not mind techie language and if you wish to monitor the progress, looks like the easiest way to sign up for this slack channel is to go to: https://omniapsslack.azurewebsites.net/ (but please resist the urge to ask “when” is this or that going to be done, or to make random suggestions - they know what they are doing). Later on, there will be many opportunities for everyone interested to contribute to testing, feedback, documentation, etc.

Edit: another good news: looks like existing RileyLink hardware (used with Loop) will work with the new drivers!

4 Likes

If there is one thing engineers and programmers can’t stand is someone breathing down their neck. I speak from experience.

Thanks for the update.

3 Likes

I can’t imagine why! Doesn’t everyone adore a backseat driver?!

4 Likes

Oh thankfully :smirk_cat: - I’m not the on,y one that’s clueless about this!

1 Like

Thanks @Bradford! And @Michel for expanding on that. I’m so glad because I just spent some time Monday arguing that it would soon be possible with the OmniPod with @EricH’s idiot endo (maternity coverage guy sucks, and is fired from being our endo). So yay for it being true!!!

2 Likes

Here is the original blog post btw:

2 Likes

No news here (or on the blog site you showed) for four months - does anyone who is more plugged in have any update on where things are now?

1 Like

There has been significant progress on the Omnipod closed loop project but it is a big project.

The communications encryption is now understood and the group has been working on “sniffing” the messages between the POD and PDM and developing a command list based on these messages.

A big challenge is that the POD cannot talk to the PDM and another device. The original idea was that you could operate in parallel with a PDM and send a temp basal to the POD (like the Medtronic Pumps). They quickly discovered that if two devices send messages to the same pod, the pod will fault (go into scream mode). From a device security perspective, this is a good feature for the POD becasue it means someone else can’t take control of your POD without your PDM, but from a hacking perspective it means more work.

They have to basically recreate all the PDM functions in the device that hosts the APS software and that device will become the PDM. This is what is taking so long - to figure out all the commands and alarms etc.

Yesterday someone on the development group got a raspberry pi to bolus so there is some functionality there, but there is a lot more to go.

I would say there is definitely momentum on the project and lots of people are contributing. I would say at this point it is a matter of when and not if, but you never know.

There is a github site set-up for this work and the current status is on the github page.

9 Likes

It is so encouraging to think that we may be able to loop on a current pump, rather than having to rely on old fossilized Medtronic pumps that are years past their expiration date.

2 Likes

You can loop with the modern Dana-RS in Europe and Asia. I think FDA approval in 2018 is expected and one day it may be available in Canada.

The Dana RS was specifically built to work with AndroidAPS.

http://www.sooil.com/eng/product/

But then again I would rather have a tubeless pump :smile:

3 Likes

Same here!

But I am also looking eagerly forward to the Dana pump :slight_smile: Two options are better than one, either is better than older pumps, and older pumps are MUCH better than none!

2 Likes

A post was split to a new topic: New JDRF partnership for closed-loop pump in Korea

@Bradford just posted about this interesting new closed loop pump project in Korea:

Incremental progress with Omnipod. Looks like the OmniAPS team is slowly being able to integrate the comms into the RileyLink controller. Slow and steady. But steady indeed. Here is a snapshot of some POD info populated into RL. BTW if anyone has any expired PODs they no longer need, please head over to the the SLACK channel to donate if you are able. https://omniaps.slack.com/

7 Likes

Really excellent news.

Does anyone know if the existing Rileylink will still work as designed with Omnipod? @dm61 was mentioning that it may need a redesign to be as effective as possible with Omnipod.

It looks like they are using the old Rileylink hardware for testing so far.

1 Like

Looks like that’s the case, same Rileylink, but with updated firmware. They’ve made excellent progress, but there is still lots of work to be done.

1 Like