telemetry?

I am looking for a robust and reliable digital telemetry link between a solar car (which may be up to a mile away, on track) and not necessarily line-of-sight. Requirements listed below:

* >1000 baud preferred (100 baud minimum)
* Simple to implement (we don't have lots of development time left for the project)
* 1 mile distance with obstructions
* Legal (I don't want to piss off the FCC :P)
* efficient (let's impose a... oh, IDK... 5W transmit power for the car? Ok? Ok.)

What I have considered / am considering:

* Other teams have used the popular 900MHz Xbee pro UART radio modules for this purpose, but the distance seems to be quite short / limited. The advantages include eese of use, as it acts as a nearly-transparent UART bridge, but it appears they may not be good enough for this application.

* 433MHz? I know lower frequencies tend to propagate further for a given power, would 433MHz be preferable over 900MHz? I can't seem to find many devices beyond small keyfobs and stuff using this band.

* 72MHz? I know this band was used for old-school RC transmitters, can I repurpose it for telemetry? Have not looked into it too much. Doesn't appear to do any sort of frequency hopping or other fancy stuff either.

* amatuer HAM / CB / radio frequencies? I think this is the most robust solution but probably requires the most effort too. I and the rest of the team would need to get licenced (although honestly I want get licenced to mess around with this stuff anyway) and we would probably need to design a system around some form of digital encoding. There appear to be a lot of different methods, but are there any that meet the throughput requirements?




sort by: active | newest | oldest
rickharris4 months ago

1 mile with obstructions will be very hard to get reliably with legal none commercial systems.

Most telemetry for cars I have seen relies on them being on a track and so they collect data as the car goes past the pits.

Most telemetry is a waste of time :-) You can't do much with it during use loss of time to make adjustments is greater then the on track losses (in my experience).

Anything else is weight and power drain that your better off without.

-max- (author)  rickharris4 months ago

While observing the "strategy crew" of another team, CalSol, They used xbee radios and like you say, only received the datastream when zooming by the pits.

ideally I would like full video telemetry showing real-time video of the road and data to accompany it. I wonder if using a good 900MHz FPV setup would work, and using the audio channels with modems on either end...

rickharris -max-4 months ago

If your racing ANY weight that isn't necessary decreases your chances of winning. ANY weight. Unless the telemetry gives you a clear advantage over the other teams (and if they have it how can it) your better off without the weight. Spend the money and effort on developing lighter body work etc.

-max- (author)  rickharris4 months ago

The team already built a carbon fiber "bathtub" for the body, and the estimated weight is looking like 340Kg. With it, we are already screwed when it comes to being competitive. This is not helped by the fact that we only have a single 10HP hub motor for a 4 wheel design powering a single wheel. Or that we are planning to use 2000mah 18650s from 2005 as we cant afford better cells. (Frankly it would be a waste of money anyway with the body design.)

This is more about getting on track and learning as much as possible. A few ounces and a watt or so is certainly not our critical point, I guess its not strictly nessesary, as such data can be logged to an SD and overlayed with a video.

However, enabling everyone on the electrical team to see how variables like battery voltage(s), current, solar charge, thermals, regen current, all relate as compared to the driving conditions as well as the driving/cruising strategy of the driver and having the ability to interact with the driver to see how changing his/her driving stratagy affects efficency would be invalueable.

Where's the weight ? In the CF body ????

-max- (author)  steveastrouk4 months ago

I don't know, I was given that figure as an estimate for the total weight of the car. The body was constructed by creating a "positive mold" with a dense fiberglass-reinforced structural foam and this had many many layers of CF applied over it. As it was the first attempt ever manufacturing with CF, it is not optimal by any means. I helped move it and it was quite heavy.

But this is far beyond the point. I can't see why some basic telemetry would be so difficult... For the time being, we might just go with a few 250mW 900MHz Xbee modules and circularly polarized antennas.

rickharris -max-4 months ago

Then go for an onboard system that collects the data for later interpretation.

honestly your entering deep deep water to be able to reliably collect transmit data AND video.

Do it on board.

Downunder35m4 months ago

Might be a bit of a long shot but since I am thinking of using something similar on a remote property it could work for you too:
Using the LoRa module you can transmit data over quite long distances.
It also supports various platforms so integration should be fairly easy.

https://www.cooking-hacks.com/blog/send-data-at-ex...

For live video there is always DJI's lighbridge technology - in case you want to see what is happening.

I had the chance to talk to a few guys doing the solar races here in the AU outback for a few years.
They stopped and got real jobs now that they are finnished with uni and never actually managed to finnish a race but still...
To overcome the problem of really no cell phone coverage and licenses these guys used a hired "satellite logger" for their communication needs.
Data was "compressed" into short text messages and then sent at regular intervals through the "phone".
If I am not fully mistaken then this is one of the modules they used during their last races:

http://www.halytech.com.au/products/dataloggers-an...

Hope this gives you some more ideas to play with, although they might be a bit more costly.

bwrussell4 months ago

Go get a 4G hotspot from a wireless carrier and just transmit the data over the internet. You might have to transmit it off the car to a base station first if you can't get the whole track inside the hotspot's range.

-max- (author)  bwrussell4 months ago

I'm afraid these 2 solutions do not meet the "simple" requirement though. The esteban team did actually create a remote server using cellular networks, doing this with 3G data. However they were not able to use it due to certain solar challenges being in remote areas of the globe.

As the team is a fresh reboot starting essentially from scratch, with no prior experience in design, The electrical lead of the team as well as I both agree that the design should be kept as simple and robust as possible for the first solar car. The goal is not to be competitive, but rather to just pass scrutineering and get on the track.

How can you beat "off the shelf" with no, or next to no, hardware and software development for "simple". Your caveat of "they were not able to use it due to certain solar challenges being in remote areas of the globe." was not in the specification you presented.

If I had to do it off grid, I'd use Iridium.

-max- (author)  steveastrouk4 months ago

I don't have any idea how I would use a 4G "hotspot" to do this. Advice would be appreciated. I am under the assumption that this would require learning about the different layers of networking, node.js, server / client scripting, javascript (although I do know the basics of js), etc etc. It all seems just a bit intimidating, at least to me.

At the very least, it does not seem as easy as simply hooking up rx/tx on a teensy 3.6 and setting up a PuTTY session, and possibly throwing some basic java / processing / python gui on top for creating visual representations of different data.

bwrussell -max-4 months ago

It's honestly pretty simple to set up something basic yourself. It's all been done before and there should be plenty of tutorials/examples out there. Most micro-controllers/micro-computers out there either have a built in or add-on networking solution. I'm not sure what your sensor system is like but something like the Seeed Studio Linkit One, which you can plug a SIM card directly into, might be just the ticket.

You could even look into using something like the Intel Edison which links up to Intel's IoT website providing built in data logging and visualization. It might not be "direct connection real-time", it's been a while since I used it, but it would run at the speed of a refresh at the slowest. It wouldn't be the best long term option but it would be a start while you develop a bespoke solution.

+1

You're wildly overthinking about it. The underlying process is moving data between "ports" at IPaddresses. It really COULD be as easy as a terminal session, though I would use SSH.

steveastrouk4 months ago

I agree with BWRussell. 4G hotspot.

-max- (author)  steveastrouk4 months ago

I, nor any other team members have much/any experience with server applications or mobile app development (as far as I know) Given the primary objective of our current solar car is to get something that just works in a rather short amount of time.

You can always learn. Now's the time to start. Its not terribly hard, the hotspot is doing the tricky stuff.

bwrussell4 months ago

You could chain a mesh network to transfer the data back, but if you have the access to all the land between the track and where you want the data to do this you're probably better off just going wireless to just off the track and then dropping a cable back to you.