In the last week there was a bit of buzz about a project by Yan Wan from the University of Texas who, at the Smart America Expo, showed a WiFi enabled drone that *ding* could provide WiFi with a range of up to 5 miles! This was presented as a great step forward in disaster communications. The various articles made it seem like one of these devices would cover a 5 mile area with WiFi.
See for yourself by the viewing the video here.
Having worked in the fields of wifi and disaster response I decided to have a bit of a closer look. Was this really such a step forward and whatsmore, how was this achieved…
Firstly let’s have a look at the hardware, what components were used to build this. The aerial platform is a DJI F650 platform. It’s a nice of the shelf platform. The payload is a Ubiquiti Nanostation. Then there is some sort of servo which rotates the Nanostation. Wether or not the rotation is automatic and slaved to the location of the station the devices connects to or not isn’t clear.
The Nanostation is a widely used WiFi AP/CPE operating in the 5GHz band. I’ve used them on different occasions and we recently deployed some when working in the Philippines with Disaster Tech Lab. They’re a functional, cheap and cheerful type of device which theoretically can provide up to 150 Mbps on a point-to-point link of up to 15 km.
So in short this project has taken two of the shelf products and bolted them together.
Now for the technical bit; The Nanostation is designed to be one end of a point-to-point link and as with all point-to-point links range/distance is achieved by narrowing the antenna beam. While an omni-directional antenna will cover a 360 degree radius it’s range is limited. A directional antenna, like the one inside the Nanostation, takes the same transmission power (TX) but squeezes it into a narrow bundle. That way the range is much further but the signal radius is much smaller. Any device outside the range of the bundle will not be able to connect. Compare it to standing in a huge dark room and switching on a lightbulb versus a flashlight and you’ll get the picture. In addition the touted range of 5 miles is really not such a big deal. Most of the off the shelf WiFi point-to-point devices have a much longer range. I can recall building a link over 40 kilometers back in 2004 (I think) using homemade gear.
So going back to the implication of this aerial platform; we have a radio device with a narrow beam mounted onto what by definition is an unstable platform. Just ask anyone who has ever build a point-to-point link using masts with a lot of wind-sway how this can kill your connectivity.
Then there is the issue of backhaul; connecting two WiFi devices is great and all but it means very little without a source of internet backhaul. As this is an airborne device using a cable is a big no-no. So a form of wireless backhaul is called for. 3G/4G/LTE? All possible but it will add to the size and weight of the payload as well as require more battery power. In addition, if 3G/4G/LTE is available would that not negate the need for an airborne wifi device? So satellite maybe? That would mean an ever bigger payload and battery drain. The last option is to consider using lots of these devices and building a WiFi mesh with one or more fixed end-points with internet connectivity. Now this would theoretically be a possibility but there are several factors to consider. using directional antennas on airborne/unstable platforms might cause dropped links. Then there is ones of the bigger drawbacks inherent to WiFi mesh networks; a drop in throughput with every hop (unless you use dual radio devices but that won’t really work here). Lastly there is the logistics and costs of building, deploying and operating a large number of these type of devices.
So considering all the above I can only conclude that it’s a case of much ado about nothing. It might look cool and get much ooh’s and aah’s from the hackathon crowd but there’s really no benefit in this product at all. The cost of the combined package comes to around $1300 (off-hand) with $1000 of that being the aerial platform. Compare that to about $150 for a length of scaffolding pipe, a roof mount kit and some clamps. The latter would also create a link with much more stability.
If you are looking at a workable application for aerial platform mounted WiFi then look at a combination of a 3G/4G/LTE backhaul device and a WiFi router with an omni directional antenna. That way you can create small bubbles of WiFi that can be flown into hard to reach areas and which are not dependent on a directional link for connectivity.
Even better rather than a means of connectivity provisioning use aerial platforms as sensor networks able to detect wifi or cellular signals transmitted by client devices (smartphones etc) pinpointing people in need.