february 2002 archives

Wednesday, February 27, 2002

At last I made the big step of giving my website a new look: two columns! Yeah, and without tables. It's all CSS. “Schtop, schtop, it's not ready yet!”. There is some minor tweaks I need to do, so do not hold it against me, yet...

Tomorrow, a rock concert by Zita Swoon, a Belgian rock band recommended by Harold. Really looking forward to seeing it!

Relying on a government-owned satellite programme for positioning, let alone its lack of positional accuracy in urban areas? Apparently my concern is shared. So, what are the alternative. Broadly speaking, there are three options: TOA, Cell ID and triangulation. Other solutions combine these options. Each option has its own postional accuracy. Yeah, you still have to wait a bit more...

Tuesday, February 26, 2002

A GPS in every car? If it were up to the chairman of Britain's Commission for Integrated Transport, the answer would be “yes”, according to “Satellite cops to patrol UK streets?”. He proposed a plan to tax car use in congestion-prone areas. Red-lining urban areas, would introduce a tax system that emphasises car use, not car ownership.

The GPS receiver compares the time a signals of at least three of the 24 satellites to determine how far away the satellites are. The location of the car is then calculated by triangulation. Its location is then compared to the polygon that describes the red-lined area to determine whether or not to charge the congestion tax.

But what about positional accuracy, because you would not want to be charged if you haven't entered the red-lined area. DoD, as operator of the GPS, now provides civil users a horizontal positioning accuracy of 36 meters, compared to 100-meter accuracy in the previous edition of the standard, which was published in 1995. Garmin and Magellan quote an accuracy of 3 to 15 meters.

It is well known, that the accuracy can fluctuate. Especially in build-up areas that are most likely to be designated as “congestion-tax zones”, this accuracy can be negatively influenced. The GPS signal is reflected off objects such as tall buildings before it reaches the receiver. This increases the travel time of the signal, thereby causing errors. Also the number of satellites visible determine the acccuracy. The more signals a GPS receiver can lock on, the better the accuracy. Buildings, and electronic interference can block signal reception, causing position errors or possibly no position reading at all.

Furthermore, Selective Availability (SA) is an intentional degradation of the signal that can be imposed by the US Department of Defense. SA was intended to prevent military adversaries from using the highly accurate GPS signals. The US government turned off SA in May 2000, which significantly improved the accuracy of civilian GPS receivers.

In the GIS community, the “Current Situation” created a buzz around SA. On September 17, 2001, the Interagency GPS Executive Board, which governs the system, posted a statement on its website by saying SA will never be reinstated. However, the Pentagon has developed the capability to jam civilian GPS signals within a specific targeted area and could easily deny the 36-meter-accuracy civilian signal, without interfering with users in other areas of the world. That would mean only military GPS receivers in planes, ships and in the hands of US special forces would work within the targeted area. But can we really rely on this statement?

Positional accuracy seems to be the topic of today, actually. I also came across the paper “Setting the Correct Accuracy Expectations for Different Location Based Services”. Red-lining is also one of the topics addressed in this paper. But there is more to positional accuracy in LBS. Let me entertain this further on a later occasion.

Monday, February 25, 2002

Yeah, well. I know... I am responsible for Multimap.com's online travel directions service, but.... I do not have a drivers licence. So this weekend, I have been chauffeured around. On Saturday, we went to Oxford after a good, long lie-in. It was great to be in Oxford again. Beautiful buildings, lots of history and of course the studenty atmosphere. Oh, how do I miss being a student... (sigh).

Especially good was the visit we paid to one of the many Blackwell's bookstores there. It is just amazing how many books they do. I had a good browse through the “Geography” section. Finally I saw the WebCartography book by Kraak and Brown. I hadn't laid my hand on it before.

Yesterday, we paid a vist to IKEA. I hadn't been there since a long time. Saw a lot of small stuff to buy, but it was not really worth queuing up for. It was a Sunday after all. Maybe call in a sicky one day and go on a Tuesday? Just kidding... From country lanes to motorway junctions, this weekend has provided me with enough field information to go back to the drawing board.

Friday, February 15, 2002

Mad about maps? Some people can really get frustrated. Imagine you run a company on Denman Street Central, Radford that has the postcode NG7 3PS. On the map it all shows up alright.

However, do you see the small road off Denman Street Central that runs parallel to Radford Boulevard? It's on the OS map and the name of the street is “Bulwer Road”. The coordinates associated with the postcode “NG7 3PS” actually describe a point that has this as its nearest road, not Denman Street Central. So what happens if you generate travel directions to the postcode “NG7 3PS”? It tells you how to drive to Bulwer Road, not to Denman Street Central!

Not much of a problem at first sight, as you would probably get to Denman Street Central and realise you have arrived at the proper location anyway. OK, of course it's a bit daft if it is for your company. Things are not all what it seems to be... Local knowledge will tell you that the road name has not been in existence for 15 years! Ooops!

It's not just Multimap.com that shows this error. Also on Streetmap.co.uk, you see the same “ghost” street. Looking for the postcode at Maporama, returns a location in the middle of a building block, but also shows a small street off Denman Street Central and so does MapQuest. This means that there is a street in the OS dataset and the NavTech data set that. The OS data set adds the name Bulwer Street to it. Would this be a case of fingerprinting?

Well, there's a street just south that is labelled “Bulwer Road”. Also Maporama and MapQuest identify this road. Whom to trust? An angry shop owner? The data? Or is the Ordnance Survey fingerprinting their database?

Spatially aware birds... Nature runs an article about a research into the brain activity of pigeons as the animals walked around a room. The hippocampus, a part of the brain that learns and remembers surroundings, is thought to be partly responsible for navigational expertise. Pigeon hippocampal cells have more complex firing patterns, and are triggered by more than simple location. Exactly what aspect of a scene they respond to remains unclear.

Another bird species that has excellent navigational expertise is the nutcracker. After hiding winter seed caches in up to 6,000 separate spots, they unearth them again over winter. Maybe scientists will be able to define which chemicals trigger this activity and then we all will be able to read maps?

Thursday, February 14, 2002

Location-based entertainment is getting closer! A few months ago, I mentioned some applications in Scandinavia and Germany and Switzerland. However, BotFighters will be launched in Ireland shortly, according to an article in Wired. Another location-based game called “X-Fire” was launched last fall in conjunction with Britain's Channel 4. Based on a popular paintball TV competition, the game attracted as many as 11,000 people in the United Kingdom to sign up for the location-based version before it was temporarily shelved after Sept. 11.

A few days ago (February 12, 2002), Navigation Technologies Corp. followed the TeleAtlas example of delivering map data in a voice-enabled format. It announced six country, six language rollout of voice data. Philips Speech Processing is leading the Europe phonetic transcriptions while the US phonetic data is being provided by E-Speech. Systems leveraging the voice data will enable users to speak directly to their navigation systems to tell them where they want to go, while the system will speak back to them, telling them how to get there using actual names every step of the way.

Already last year (March 12, 2001), TeleAtlas BV sent out a press release that it had increased its “share power” in voice-operated geographical applications, buying all the remaining shares of the Belgian voice technology business Phonetic Topographics NV. Its main product, “MultiNet”, now contains phonetic transcriptions for several regions and languages, enabling text-to-speech and Automatic Speech Recognition in a variety of applications.

Multimap.com has increased its delivery of online travel directions by 50% over the last couple of days! Demand seems to be still increasing. Also MSN Hotmail features “Routeplanner” again in its Top Ten Links. By the way, have I mentioned before, that MSN has come up with its own online travel directions service: MSN MapPoint.NET? Well, on MSN.nl, the service is still provided by Locatienet.

Thursday, February 07, 2002

“webmapper.net - maps for the oyayubizoku (the thumb tribe)”

Sunday, February 03, 2002

Haven't updated you for a while. Sorry 'bout that. Last weekend Harold and Anne came to visit me here in London. It was really great to see them again. Due to the weather we could not do as much as we would have liked, but we were exhausted nevertheless...

This week was rather exciting. On Tuesday, I delivered a presentation at InfoTech's on their online routing software RouteXNetServer. The audience consisted of InfoTech resellers in Europe. It was very interesting meeting people from Portugal, Finland, Norway. It almost felt like being back at ITC again :-)

Hope this week will be a little less stressfull..