3.6GHz 802.11y for Competitions

[magnets]

People brought this up in the Einstein 2012 thread, and they realized three things.

1. 3.6GHz is a licensed channel, so first would have to pay.
2. The problems (excluding those of Einstein 2012) aren't caused by interfering signals.
3. If implemented properly, normal 802.11n should work fine, even with lots of other networks happening at the same time. Instead of making it more complicated to fix the problem, FIRST should resolve the current issue.

Overall, the 3.6GHz solution requires too much work (money, new radios...) in order for something that may not be the right solution, may make the situation worse, or could be resolved another way.

That being said, using a different frequency would be great. The old IFI radios were 900MHz, and had a pretty decent bandwidth, more than what we need for control data. The camera data could still be sent over wifi, and they even make axis cams with built in wifi.

[techhelpbb]

Originally Posted by magnets

3. If implemented properly, normal 802.11n should work fine, even with lots of other networks happening at the same time. Instead of making it more complicated to fix the problem, FIRST should resolve the current issue.

Even if FIRST reduces each field to a single channel and makes the field/robot system more stable the effect of reusing a channel will be degraded performance with one or more network.

Once the other network is out of FIRST's control the risk of channel monopolization is high.

As we have both stated putting the show stopping robot communication on a different channel is possible. That other channel should be easier to diagnose without robots of various different needs placing demands on it. The field systems are common and required for all fielded robots. TCP/IP and video is not required but nice to have.

On the topic of 3.6GHz it only perpetuates binding the necessary field communications to the extra optional traffic.
It would be far cheaper to use RS232/TTL serial to RF adapters for the field traffic.
Licensing issues may apply to both options.

Course we could try IRDA for the fields and make the fields more RF immune.

[EricWilliams]

Originally Posted by magnets

1. 3.6GHz is a licensed channel, so first would have to pay.
2. The problems (excluding those of Einstein 2012) aren't caused by interfering signals.
3. If implemented properly, normal 802.11n should work fine, even with lots of other networks happening at the same time. Instead of making it more complicated to fix the problem, FIRST should resolve the current issue.

The fact that it's licensed is stated in the proposal. Many telecommunication technologies require an FCC license for use - the reason these bands/services are licensed is to protect the users from interfering signals. As that's the point of this proposal, I feel it's an acceptable cost.

From the FCC site (http://wireless.fcc.gov/services/ind...&id=3650_3700), it appears the filing fee would be $60 + $150/Call Sign - at worst, this is less than $2000 (somewhere between $50 and $100 per team) These events cost money - the A/V budget alone for some events would most likely be mind boggling. I think that's a reasonable price to pay to ensure that the command & control network remains isolated and secure.

Originally Posted by magnets

Overall, the 3.6GHz solution requires too much work (money, new radios...) in order for something that may not be the right solution, may make the situation worse, or could be resolved another way.

I don't agree, or at least don't find enough evidence to support that statement in your comment to summarily dismiss the idea.

As for going back to some type of custom hardware/radio: that completely undermines the biggest benefit of going away from the 900MHz radios - that teams can use standard COTS equipment for development/demo/practice/off-season events. I'm suggesting simply swapping out the bridge before you bring the robot out onto the field.

[techhelpbb]

Originally Posted by EricWilliams

As for going back to some type of custom hardware/radio: that completely undermines the biggest benefit of going aay from the 900MHz radios - that teams can use standard COTS equipment for development/demo/practice/off-season events. I'm suggesting simply swapping out the bridge before you bring the robot out onto the field.

When I helped propose a 2015 control system I got around this by making a module that accepted radio modules of various frequencies. FIRST could send teams ISM band modules (often unlicensed) with valid frequencies for the team's nation of origin. If you really wanted that module could do TCP/IP for home use of an AP or base band serial. After all that module had an Ethernet for field side. It could work with existing fields and FMS. The module was called the Turtle. Slow and steady wins the race and it is hardened.

If the teams wanted WiFi FIRST could let them provide hardware or restrict the hardware to choices they prefer.

Turtle did not advocate a specific frequency it would have allowed FIRST to move around.

This proposal was not accepted.

I think Asus is providing a WiFi USB radio for RoboRIO.
Upside - less involved than the D-Link AP and hopefully cheaper.
Downsides - still wraps baseband serial in TCP/IP giving up control and risks RoboRIO busy state holding up radio.