About time...

[jducoeur]

It's always neat to see one of my blue-sky ideas come to fruition. (A little frustrating to see that they went and *patented* the damned thing, but that's what I get for not going to the effort myself. If I ever care to invalidate the patent, I can investigate whether I ever really talked about it publically.)

Anyway: one of the points I've been making for, oh, probably six or seven years now is the killer app for Internet to the car: a truly smart navigation system, with up-to-the-minute information. Well, it looks like it's finally happening: the Dash Express is almost exactly the device I designed all those years ago, and have talked about occasionally since then. It's a standard GPS system, but is Internet-connected to their servers. This is key in two ways. First, it uses current traffic information to build its route suggestions -- rather than simply going for the *shortest* route, it will automatically take you around traffic. Second, it uploads and anonymizes the GPS position and speeds from its users to build up seriously accurate traffic info. Once enough people in a given area have the system, it ought to wind up with the most accurate traffic picture out there.

They're apparently in alpha now, but the reviews are starting to trickle in, and it appears to be *very* nearly my ideal device. It fails on only one front: I would have preferred to separate the hardware and software equations. If the front end device was a general-purpose Internet terminal, with the Dash navigation software as just a single app on it, they would have opened up a much bigger (if slightly riskier) opportunity for themselves, to become the cornerstone of a whole new industry. I may yet write to them and suggest that...

Comments

[osewalrus.livejournal.com]

Does this use the satellite feed or some other spectrum connectivity.

As anyone who has tried to maintain his or her mobile phone connection while driving knows, mobile connectivity is a major pain in the neck due to hand off issues and the "concrete canyons" problem. OTOH, satellite probably can work here, because the info is relatively low bandwidth and latency tolerant.

I wonder if any of the cellular companies will try to replicate terrestrially with the AWS spectrum they just got?

[jducoeur]

Does this use the satellite feed or some other spectrum connectivity.

I don't know, and I'm quite curious about that. For all I know, they could even be using something WiMax-ish, if they're really focusing on municipal areas with enough density. There are so many options for wireless broadband now, and none of them are perfect, so it would be interesting to know which tradeoffs they chose.

OTOH, satellite probably can work here, because the info is relatively low bandwidth and latency tolerant.

Does satellite provide 2-way at this point? (Entirely possible -- I haven't been paying attention to that technology for a while.) But yes: this app ought to be able to cope with latencies and bandwidths that would be out of the question for many. (Especially if it's doing smart lookahead and caching.)

I wonder if any of the cellular companies will try to replicate terrestrially with the AWS spectrum they just got?

Good question. I wonder if my father has any info on that. (He's the broadband expert at this point -- he and his wife are basically professional futurists for the home-broadband industry, telling the cable operators what they need to do in order to stay competitive...)

[oakleaf-mirror.livejournal.com]

I'm sceptical that this will produce useful traffic info any time soon, from the feedback. GPS nav systems are still a niche market, and to get one that specifically reports back with the traffic data will be rare, for some time. You could be stuck in traffic, yet appear to be the only car on that road for the silence of the other cars.

My GPS does get and use traffic information. It's a one way link with the data coming in via XM satelight. The GPS does use that info to automatically route around traffic.

Tom Tom makes a GPS that connects to the internet via GPRS data through your mobile phone. The unit also functions as a handsfree set for your phone via Bluetooth. It will use your phone's data connection to download traffic data from the net, and auto route around that. I don't recall that they do much else with the link, though.

[alexx-kay.livejournal.com]

You could be stuck in traffic, yet appear to be the only car on that road for the silence of the other cars.

Well, you could infer a fair amount with clever algorithms. If the car is moving in stops and starts, at about 5 MPH, on a major highway, that's probably an indicator of heavy traffic.

If the company is allowed to store individual driving data over long times, you can get even cleverer. "90% of the time when on this road, this driver goes between 60 and 70 MPH. Today they are doing 30, so there is probably some sort of traffic problem." Of course, storing that sort of data would have obvious privacy issues...

Now, none of this lets you know about a traffic jam that has *none* of your cars involved. But at least once one of your customers is trapped, there's a chance that you can avoid sending another one into the same mess.

[jducoeur]

Now, none of this lets you know about a traffic jam that has *none* of your cars involved. But at least once one of your customers is trapped, there's a chance that you can avoid sending another one into the same mess.

Correct. It definitely can't be perfect, but you can do a lot with smart heuristics. For extra credit, they should be monitoring how much traffic density seems to be necessary in order to cause a road to back up, and using that to avoid sending all of their customers along the same alternate routes. (I don't think they're doing that yet; instead, they offer several alternatives at a time, and presumably expect different customers to choose different alternatives...)

[jducoeur]

I'm sceptical that this will produce useful traffic info any time soon, from the feedback.

Their claim (that I saw somewhere) is that they need 3000 customers in a given municipal area to get reasonably good coverage of the main roads; that sounds plausible to me. One of the more interesting questions is whether they've inked deals with existing traffic-monitoring companies (eg, Smartraveler) to get "bootstrap" data to start off with. I believe they'll need that in order to get off the ground for the first 6-12 months in an area.

Interesting news about the other systems that are incorporating live traffic data. Hadn't known that was percolating in so much already; thanks!