Pandemic prevention and management is an interesting territory for IoT. I’ve seen different projects and solutions made in emergency to help during the Covid-19 crisis like some connected button distributed in hospitals to get assistance for going to toilets. That’s an interesting use case for IoT where you can deploy quickly a solution (20K have been produced in this example) and make it working in an existing environment (an hospital) with no integration complexity with network or other existing infrastructures.
My favorite example is the Us health Weather Map reporting the human temperature sensor information across the USA. I love this use-case because it works on prevention: detecting the rise of the pandemic, show its movement and scale. The current problem is that these data comes from high-end thermometers, using Bluetooth and the number of device is limited. The other problem are the personal information attached to these device.
This is where LPWAN and Sigfox in particular can be used to propose a innovative solution I’m going to detail a bit.
The purpose is to have human thermal sensor massively deployed to get precise insight about country wide health / world wide health. I mean this idea is basically to distribute such sensors to the whole population to anticipate and correctly manage a pandemic crisis with real time data.
For this, we need a solution able to communicate, respecting the privacy of people at low cost to be massively deployed.
A basic thermometer costs only 5€ on Amazon like this one.
For this price we have an electronic board, a battery and a temperature sensor with the right precision.
What we need to have, on top of it, is a radio module able to communicate, a small PCB antenna and a little piece of software able to report a simple trigger, eventually no bit of data. This reported information would just be the following: “I’ve detected a human temperature over 38°c”. And basically we need no payload data for this as getting a message from the device means this information.
The network will add the necessary information to know the location where this has been reported. This is obtained by the triangulation capability of any LPWAN network. The precision is not important and a benefit to the user privacy. Basically with 1km of precision we are good to track and manage a pandemic, in my point of view.
The good news is that this technical solution exists: it’s the Sigfox Ultra-Low cost design and its integration in an existing design like to one above could be made for a price of 0.60€ only.
As an example in France we have 29M of families, equipping all of them could be a cost around 150,000,000M€, surely much less due to the scale factor. This solution could report precious, real time, information during many years.
Such LPWAN solution allows to have no technology complexity for the end-user as there is no link with any smartphone or home equipment. You use that connected thermometer as any non connected thermometer. You don’t need to have any registration, so your privacy is really good. The thermometer can be distributed randomly we no personal information attached on it.
The only other point to consider is the connectivity cost, but I’m quite sure governments can have a really good deal for a such topic regarding the really small quantity of data that Sigfox would have to manage and the scale. Then, the only limit is the Sigfox IDs at 32b … isn’t it ?
Let’s tune this idea
After some hours/days thinking about this idea, I’m less and less confident that a product is the center of this solution. Basically creating a product, sponsored by states will create multiple standards and will be against a competition in the market. At the end, why the consumer would use the “state thermometer” instead of another one he has or a more modern model?
Taking into account this point, my feeling is more going to create a consortium and a standard sponsored by governments.
We need to have in this consortium the following partners:
- A network operator to propose and manage the connectivity (Sigfox)
- One or multiple chip manufacturer able to propose a ref design really easy to implement at low cost (let say Cmostek)
- A company able to manage the data, store them with a high level of security and in a distributed way (I will explain why). Then able to analyse and publish them. (I have some idea here for a company 😉 )
- Government & public scientists representative for decision and governance. For financing.
With such combination of actors, any industrial making thermometer in the world, could access the technology for free or a couple of cents or getting payed for this. They would ask for the communication solution to the consortium and get it. This way, for the states, instead of paying for the whole thermometer they would only have to finance a smaller part.
This radio part would be already flashed with the right firmware and IDs so the thermometer manufacturer would have nothing to know about the related technologies. That just a device to power and where you send a trigger. This should basically cost less than 0.60€ per module for large volume around 10M pieces. This can be optimized as the cost of the Quartz is important in the solution and regarding the number of messages exchanged I think you can reduce the spec without really impacting the network capabilities. I’m investigating a bit that direction to save about 0.10€ more.
Having a low cost communication module is important but if we want the manufacturer to add the feature in a low-cost, mass market, thermometer we also need to give them a financial incentive. For this I assume we can pay them the price of the module plus a fee to support the initial investment and improve the margin. We can have a symbolic price of 1€ including 0.60€ of the previously seen module + 0.40€ of incentive per thermometer.
The network operator provides the communication solution the need is is really specific compare to usual communication as each of the devices will have a communication only 3-10 times per year. So I assume the price can be tuned to reach a really low price. I’ll take an assumption that seams to be realistic for hundreds of million of devices world wide. Let say 0.10€ / device per year.
The role of the company managing data is key. This company is a trusted third party, audited by the state members. The data also have to be distributed to be able to be audited by different governments. The data have to be highly secured. In my point of view, even if we have no personal data in that solution (no possible re identification) this dataset could be manipulated: you want to stop an economy? manipulate these data. You want to close up your citizen? manipulate these data. So for these reasons this entity is a key element. This entity supports the run costs and make the analysis. It’s cost should be around 0.20€ per year in my point of view. Price is a question of scale as this part own a large part of fix costs but is service oriented with a subscription approach.
So basically we have a solution with a total cost composed the following way and financed by governments or international agency like OMS:
- 1.00€ per thermometer.
- 0,30€ per active thermometer per year.
In my point of view, this approach is efficient we have no over cost for the end-user, no change in the way they are using the product. There is a limited investment for the governments to access this data and the service part paid yearly can be adapted by making a compromise between money spent and precision expected.
Let summarize this:
Open this idea to larger area
Why only accepting data from Sigfox ? Many other technology of connected thermometer exists. Some are really low cost – I discussed with a company making a sensor working with a smartphone, only connected through the JACK audio cable and cost quite nothing. Many people have a Bluetooth thermometer.
So let’s make a standard and open API where thermometer providers could report data with a simple interface respecting the conditions previously defined : only a trigger, no personal information and a location precision of 1km2 even if you can get something better.
This API only have one consequence: how to trust the data and ensure the reports are valid. Around this question I assume we can fin interesting question around identification & anonymization.
Certify the data
Another interesting question is how to certify the data. Basically such kind of data is really sensitive in a geopolitical point of view. The trust tiers is one response but usually a such tiers depends on one of the country. We can imagine to use multiple tiers but I think the right modern way is to use a blockchain. That’s a really interesting blockchain use-case 😉 where the different states part of the consortium will participate to the data certification. The network can provide a proof of location and the data can be registered in a blockchain.
Proof of Technology
The starting point of this idea is the creation of a such thermometer, I’ve post a fully blog post on how to hack an existing thermometer to make it a smart thermometer. Here is a short video demonstrating it.
If someone loves the idea, I would be please to help developing this ; just let me know.