Connecting a moving vehicle to a Wifi (wireless internet connection) hotspot is harder than it would appear to be. We did our original tests using a normal wireless router. I held the PDA in my hand, connected to the hotspot and slowly had the bus drive away. It stayed connected for about 200 feet and it seemed like adequate time to perform an ftp.
I then automated the who process in the application. Using a separate thread, I would check every second to see if WiFi was available. If it was, I would then generate a ping to see if my server was up. If that was up, I would initiate a file transfer from the buses. That was the theory.
We then went and bought some big honking Hawking antennas, put them up, and expected all sorts of communications with the buses. It didn't happen.
The first antenna (pictured above -- located on the balcony of a jewellery store) would connect to the buses that were parked. Ironically, a bus with our equipment is parked in the picture. It wouldn't pick up quick enough for a moving target.
The second one, was a super big antenna high up on the Gucci store near Parliament.
When we did a spectrum survey, we found that there were a gazillion hotspots, and they all interfered with each other, including ours. Our prime signal strength was the area behind the fence in the pic. With the big antenna on the Gucci store, we found that we were giving the cruise ships in the harbour excellent coverage, but the street below was getting minimal signal.
We had to swap out the omni-directional antennas. Yesterday we put in the directional horns pictured above. We are moving to static IP addresses and a meshNet topology. I will let you know how that works.