FlyingPolarBear

I am looking into the best antenna for sending test data from our autonomous vehicle (www.autonvs.com), with a reliable omni range of approx 1 mile (within visual range). Which 2.4Ghz antenna products create a most reliable signal for a moving vehicle? Currently, I am looking at the SMC DI105 and DLink ANT24-1500.

Thanks ~ Dan

Thorn

Dan,

You're looking at two different types of antenna. The SMC DI105 is a directional antenna. This means that the signal travels in a very tight beam only in the direction that the antenna is pointed. Since the beamwidth on the DI105 is 50º wide, then if the antenna is pointed off-center by more than 25º, then the signal will drop off sharply, to a level where no signal will be received. Essentially, this antenna is only good if the vehicle can point the antenna back toward the area where it wants to send the data, and keep the antenna continously pointed at the right spot. Since this applies in three dimensions, you're talking some sort of powered gimbal mount.

It may be good for the the end where the telemetry is being monitored, if someone can manually aim it (and continuiously re-aim it) at the vehicle during the Challenge.

As far as the DLink ANT24-1500 goes, this may work, but you're probably going to have difficulty with it, as it's pattern is very flat. First you need to understand how omni patterns work. Look at the third post here. The DLink ANT24-1500 only has a 5º vertical beam. That means anything more than 5º above or below the plane of the antenna will have a hard time getting a signal.

Considering what I've seen of the terrain the Challange is based on, I'd recommend the following: A 5 to 8 dBi gain omni antenna on the vehicle, and an omni of no more than 12dBi gain at the monitoring station. Your signals doesn't have to punch through much in the way of foliage, but you do have to go up and down some hills. All in all, a mid-level compromise between pattern and gain should give you enough of a spherical pattern at each end to keep the communications open between the vehicle and the monitoring end.

__________________
Thorn
"Read Altas Shrugged. Compare it to today. Repeat as necessary"

kleptophobiac

I would go lower than that level of gain... 6 dBi turns into 9 dBi when carmounted to a roof. The roof "mirrors" the antenna and effectively doubles its length. I would think your best bet is to emit a very strong signal and use a low gain antenna that would not squish your pattern too small (vertically). I'd try to have a full watt going to the radiating element, and just use a vertically mounted dipole or 8" omni (approx 3db)

Would it be possible to have somebody hold the base station antenna constantly pointed towards the vehicle? If not, I would suggest a dipole with a corner reflector behind it to give a very wide pattern towards the vehicle's test area. You could also use another omni at about the 6 dBi level at the base station. Using a full watt there would be good too.

This level of power should be way overkill, but at a competition you don't want anything to fail; the power will help multipathing and field strength in a situation where a loss of connection would be bad. Polluting somebody else's RF environment? Tough. I don't think DARPA would mind too much.

If you want to have a slightly more elegant solution that doesn't require an amp, the senao 200mW cards pack a whallop with a fairly sensitive receiver.

I see that you are using a RADAR setup on your vehicle... be careful with that thing. It can overload sensitive receivers like nothing else. I learned that first hand while on a boat . Please try hard to keep the antenna in a zone of minimal field strength from the radar.

Actually... considering all of this... you'd be better off with a different communications system. How much data do you need to move? If you just need basic things (like various temps, pressures, speeds, location, etc) you could get away with something even as slow as a 1200 bps link, which should be easily attainable. A HAM radio rig running a packet station could provide the data stream you need, with far more coverage area and at frequencies where LOS would almost be a non-issue. 2m band packet can easily go 50 miles.

FlyingPolarBear

Thanks for the info. I should note that the WiFi system will be used prior to the race for testing and fine-tuning purposes, but not during the actual event. I'm particularly curious whether we will need to use a higher power access point or amplifier, or if we can get fairly reliable performance within visual range with the consumer variety low power access points with replacement antennas.

Dan

kleptophobiac

a pair of WRT54G's is really nice because of the relatively high TX power (up to 254mW with the sveasoft alchemy firmware) and the easy to get RP-TNC connectors and its flexibility.

If this is a non mission critical link, then by all means go 802.11b. It's cheap, extremely fast (for telemetry that is) and easily screwed around with.

How terrible is the terrain? The site isn't too helpful in that respect. If it is just the light hills in the picture on the front page, then a consumer grade 802.11b rig should work just fine.

FlyingPolarBear

I like the WRT54G approach with the enhanced firmware. This means fewer parts to obtain the needed power. I am not familiar in using the Linksys router as a client. How is it set up in this manner?

What range should we expect with this kind of setup (with or without replacement antennas), to ensure a strong signal with minor obstacles such as trees or cacti? If we can get a very reliable connection within 1/2 to 1 mile visual range that should be ok.

Thanks ~ Dan

kleptophobiac

If you go the 1/4 watt WRT54G way, I'd suggest to mount some heatsinks on the amplifier chips in them, just as a precaution.

You'll want to use an external antenna on both ends *for sure* to get that kind of range.

On flat ground, with that much power, and a relatively weak omni on each end (<=6dbi) I have achieved about one mile. If you have a directional antenna on the base station side and can get someone to aim it for you, I'm sure you'd get a very reliable link at that range. The only thing I'd worry about is a steep grade screwing up your polarization, but with somebody aiming for you that problem would be eliminated.

EDIT:
With the hacked firmware you can set one to be a normal AP (set to b-only) and the other as an AP client. Tell it the MAC address of the regular AP, along with channel and ssid, and all should be well.

I would put the unit acting as AP in the vehicle, and the client into a tech's hands. You'll be able to monitor signal strength far more easily this way.

agentgrn

I've been giving this some thought, and I have to ask...just how much bandwidth do you really need for this application? 1 mile isn't unheard of in a point-to-point environment, but it's going to be substantially more difficult w/ a moving target.

__________________
-A.G.-

FlyingPolarBear

We project the bandwidth required no more than 2 - 3 Mbps.

Klepto, you mentioned the ability to monitor the signal. This would be a nice feature if the router provides this information, so it can be retrieved by the PC in a straightforward manner. How would this be achieved?

kleptophobiac

The WRT54G with modified firmware has a status page with signal and noise information. I don't have the unit with me right now, so I can't verify this.... but here:

It has to get the number from somewhere, and since it runs the linux kernel it probably has a nice utility that talks to the device driver to get the signal information. I vaguely recall a huge utility called "wl" that is on the WRT54G that does everything from do active scanning to reporting results, to association. You could probably write a shell script with a curses interface that would provide a quick means of using the app (little to no typing)

I say do this on the client because I'm certain that you get signal strength information when associated to an AP, I'm not so certain about an AP giving client statistics.

2-3 mbps (real data throughput) will be a little tough on a mobile link like that... good luck!

f0urtyfive

The wrt54G can go up to 254 mw now? News to me, I'm pretty sure mine only goes up to somewhere around 84

wrzwaldo

And if it does I would sure like to see it on an analyzer. I wonder if it would look like this one?

kleptophobiac

I would love to show you, but I don't have a spectrum analyzer, just a 300mhz scope.

Anybody want to donate one?!?