Hi all,

Think of a school environment, where both closed classrooms and open play grounds, both will be available.

We have hundreds of client devices, which we dont want to communicate with each other, but only to a parent network, who will be sending data in say tens of bytes every 5-15 min interval. No heavy loads from one single client.

So now, which technology and topology should I choose for the above environment?

The tech has to be reliable and fail-proof, while cutting down deployment costs too.

Please suggest, I am so confused!

 

Welcome to AAC.

First of all you need to choose a physical layer technology. There are particular trade-offs inherent in the various options and you haven’t provided enough information to work out the answer yet.

The best way to start is to describe the actual application. What is the problem this will solve? What are the constraints that are already in place? What is the budget?

 

Welcome to AAC.

First of all you need to choose a physical layer technology. There are particular trade-offs inherent in the various options and you haven’t provided enough information to work out the answer yet.

The best way to start is to describe the actual application. What is the problem this will solve? What are the constraints that are already in place? What is the budget?

Thanks for your reply!
So here is the problem and application.
The crimes against children are going high, and the idea is to give them a safety device that would give a SOS alert when they need it. When at school, when the SOS button triggered, the same will be notified both to school and parents through the local network in school and then through internet. While they are out of school, way back home, the device will have NB-IoT chip, that will send the GPS location to the concerned people, when SOS is triggered.
There is no such safety devices for the children out there.

 

I would say there are plenty of devices with GPS localization and panic/alarm buttons on them, the main topic would be the cost not the technology.
https://gpstrackershop.com/gps-tracker-sos/

 

I would say there are plenty of devices with GPS localization and panic/alarm buttons on them, the main topic would be the cost not the technology.

I too agree, but the problem is,
One, GPS doesnt work indoors. So safety cant be assured, just with the fact that, last received location was near the school.
two, they drain batteries like something, because they are generic GSM modules that are power hungry. We cant keep them ON, all through the day. Then, with a small lipo battery, you will have to recharge them everyday, which no children nor parent will do.

correct me, if anything.

 

From what they claim the battery life on most products is at least 7 days. My suggestion is to put together a set of requirements to at least have a starting point:
- Battery life
- Accuracy
- Price
etc.

 

You have undertaken an. extremely difficult and high stakes project. A failure of the device could have grave consequences.

I am not sure I would be willing to do this, but were I doing it one of the very first things I would work on is expectation management. This is not a technical thing, but my experience demands I mention it.

Be absolutely sure that any functionality proposed, discussed, or even mentioned by you or even someone who has no direct connection, is put in proper context. A casual mention can easily become a guarantee in the mind of the listener only to haunt you later when the speculative idea proves impractical.

Written communications, such as progress reports including reminders about what your system will actually do, both at launch and eventually; a schedule for deployment with clear dates or clear statements that dates are not yet known; and answers to questions you have been asked privately by two or more people because that means others have the question too.

OK, that said, my suggestion is to avoid attempting to building your own device as much as possible—particularly for the 1.0 release version. Make a list of the mandatory functional requirements and plan to deliver those and nothing else. This is a very, very complex and fidgety project.

The first place I would look would be someplace like Particle or Blynk. Particle offers hardware as well, and both have fleet management consoles for the devices, OTA updates (you can’t live without it), and airtime. As far as non-GPS localization, you can use a company like Skyhook that provides a suite of location services including WiFi-based.

Frankly, I think you need to work out your high level architecture before you choose a communications technology. Using a service provider for airtime and fleet management is the only reasonable way to manage hundreds of devices. Using APIs from one or more providers for things like location services and messaging (see: Twilio) is the only way I can see to make this scalable and sustainable.

So, high level architecture first, the technologies to fit the subsystems the architecture requires second. That’s my advice.

 

And also while designing any device, spend time questioning your assumptions.

For example, I don’t see a need for GPS to be active until the panic button is pressed. Thus it will not draw any power until it is needed.

Secondly, I question your assertion that GPS won’t work inside. Why do you believe that (it does work inside).

Plus, as Ya’akov states this application has a significant downside if it doesn’t work when activated. Hence your design should be fail-proof. Which is more complicated than a simple push button circuit. He has many other good points as well.

 

Giving every kid a button he can press to cause a panic at his school, his home, the police and who knows who else. What could possibly go wrong?

 

Have you looked into personal safety alarms?

 

Giving every kid a button he can press to cause a panic at his school, his home, the police and who knows who else. What could possibly go wrong?

Abuse of that functionality is always an issue which to me seems rather difficult to tackle even more than the capabilities of the device. Each kid with its own "give me some attention" button, or let's press this kids button to get him into trouble.