Choosing between NB-IoT and LTE-M

Hello,

I'm building a mobile boat tracker with Arduino, which should be able to send GPS position and some sensor data over a mobile connection. I do not want to use GSM, since the GSM network (as well as the 3G network) will be completely shut down in a couple of years here in Sweden. So I'm looking at LTE based solutions instead, and preferably either of the two IoT standards, NB-IoT or LTE-M. But which one should I choose?

Situation is this: NB-IoT is sufficient for me in terms of bandwidth, and I know NB-IoT signals reach further than LTE-M, which makes a difference in how far out from the coast there is coverage. NB-IoT also consumes less power than LTE-M, which I appreciate since it's obviously a battery-powered setup. However, out of the four major mobile network operators here in Sweden, only one (Telia) has a nationwide extensive NB-IoT network, while the other three only provide spotwise NB-IoT coverage, or none at all at the moment (although it's on the rise). Everyone has an LTE-M network on the other hand.

I've figured the best solution is to order an IoT SIM from a third-party provider with international roaming agreements - that way you can benefit from the combined coverage of all the operators' networks, and also be assured that the tracker does not lose connection should it end up in Danish or Norwegian waters.

So the question really is: What is better - A roaming SIM card with access to several LTE-M networks, or a roaming SIM card with access to a single NB-IoT network? (I'm actually a little unsure whether there are roaming agreements for NB-IoT.) I might also add that I'm aware that NB-IoT does not support handover, but since no time-critical data will be transferred in this case, I was thinking it doesn't matter. I'm completely new to IoT though, and I would appreciate some good thoughts on this. For me, as long as I get the data, not even a 10 min delay would matter that much.

All the best

NB-IoT certainly doesn't support handover because it's irrelevant for the type of infrequent short messages that it's aimed at. From what I recall (having been in the standards discussions at the beginning) it ought to support roaming because some of the use cases involve tracking assets between countries, but I could be wrong. I suspect the main issue could be coverage of the areas that you want since cellular systems are generally planned for land cover not oceans.

Would the AIS system be a possibility for what you want?

To begin with, AIS wouldn't be the solution because I want to be able to track the boat from wherever I am in the world. It will mostly operate in coastal waters so I'm not too concerned about the "oceans". Of course, if I wanted to go full enterprise solution I would get myself a satellite connection, but that is far beyond the scope of this project :slight_smile: I also want to be able to read other sensor data, not just position, but all kinds of fun stuff that I can do with an Arduino on a boat. And I don't think that's possible with AIS either.

My idea was to have SD card logging of position/sensor data, so that if the boat happens to go out of coverage for a while, it can still save the data and transmit it later once it gets back into coverage. But naturally, I'd prefer if those outages were as small and rare as possible.

What do you think about that idea, you who seem to be a veteran in the business? Would a single NB-IoT network do the trick, or would I be better off with access to multiple LTE-M networks? This would also mean that occasionally, there might be a larger load of data to send. Could that cause problems with NB-IoT? What would happen if it would get cut off in a transmission when changing mast?

@jhaine : Ok, a little update. I've read up a bit and I get your idea about AIS now, it was actually a more plausible than I first thought. However, it's still limited to just position and a few basic ship data, I can't transfer custom Arduino sensor data (battery monitoring, burglar alarm, etc).

So, NB-IoT or LTE-M it is. But it seems that the range difference isn't as big as I first was told. Here's one excellent article that got me leaning back towards LTE-M , seeing the small differences in coverage/range and power consumption, with LTE-M even outperforming NB-IoT in certain conditions (check the conclusion part for a quick summary). This made me very happy, since it will probably save me a bit a of trouble. Unless you, or anyone else here have got any other input that I should consider, I think I'll aim for LTE-M.

Well I'm not entirely sure that I agree with the conclusions of that report - for example they assumed a poorer terminal receiver noise factor for NBIoT which I don't think is justified. But given that in the EU NB-IoT is less deployed and your use case requires roaming then LTE-M is probably the best choice and modules are commonly dual-mode anyway.

I feel duty bound to recommend this module: SARA-R500E module

I checked with the paper they referenced, and although much of it is Greek to me, I can't find that they make that assumption there, but the noise figures rather seem to refer to base station/client NF (eNB/UE). Anyway, I still take it that the differences in coverage/MCL are not that huge, even if LTE-M outperforming NB-IoT may be an exaggeration.

Recommendation taken, but I'm a bit sceptical to eSIM. Don't know how the situation is on the other side of the pond, but here in Sweden at least eSIM is still quite a special product that isn't offered with every subscription, and I don't want to limit my alternatives.

I had been looking at the SIM7000 before, because I've figured it already has good Arduino support and an integrated GNSS receiver. But is there any particular reason that you are recommending the SARA-R500E?

Just sentimental - I ran the early NB-IoT programme at u-blox where we made the very first modules for the NB-IoT trials in the UK.

UE = module or terminal in 3GPP speak.

Yeah, I just meant that the 9dB/5dB and 5db/3dB noise figures in the article doesn't seem to represent NB-IoT and LTE-M in the original paper, but rather eNB and UE. But I'm probably misinterpreting something.

That's cool! But if you don't mind, I think I'll go with the SIM7000 in that case :slight_smile: Perhaps when I've gotten the hang of everything in the future, I can start comparing LTE-M with NB-IoT on the same board.

This topic was automatically closed 180 days after the last reply. New replies are no longer allowed.