Hi David (Official Rep)
Last night I repeated my first experiment, which led to these comments: Note that the original experiment was before the last update.
The route is from Wendover, Buckinghamshire to Chipping Norton, Oxfordshire via the Oxford ring road.
So going back to them after the last update, there is some good news:
When will CoPilot premium be able to choose a sensible route between two points? Not even any of the alternative routes make any sense.
One of the alternative routes now makes sense
Why does CoPilot premium try to route me via footpaths when I am in a car?
It’s stopped doing this
Why does CoPilot premium try to route me through the centre of a town when there is a bypass that is both quicker and shorter?
It’s stopped doing this, there were three instances of this – all are now fixed
Why does CoPilot premium tell me (repeatedly) to keep left when the only thing it can be referring to is a large lay-by next to the dual carriageway that is full of parked lorries? At 70 mph this could end in a fatal accident.
It’s stopped doing this
Why does CoPilot premium try to route me down dead ends?
It’s stopped doing this
Why does CoPilot premium warn me about safety cameras that don't exist?
All the safety cameras on this route were now reported correctly
Re-routing appears to have improved too.
So the last update is going in the right direction. It hasn’t arrived but at least it’s the right direction.
There was one thing I did differently.
First a few thoughts about routing:
Calculating the shortest route between two points is easy, calculating the quickest is far more complex. To calculate the quickest we need to know the road speeds and bearing in mind the old computer adage “garbage in, garbage out", to get an accurate time estimation we need accurate road speeds. CoPilot asks you to input what you feel is your average road speed for various road types. There may be some validity in this when it comes to motorways, but for primary, secondary and unclassified roads it probably does more harm than good:
- Most people couldn’t accurately estimate this if they tried
- It varies so much from route to route that any average is likely to be very inaccurate anyway
There is a better source of data though – speed limits. Speed limits, in the UK at least, tend to reflect road conditions and most people drive close to the speed limit. So to get an average speed all we need to do is look at each speed zone on the route, work out the time taken for that section and then add them all up. If we want to make it even more accurate we could factor in an adjustment for each junction, roundabout etc. I know the actual algorithm would have to be more complex than this, but you get where i am going.
The question is, does CoPilot use any of this data when it calculates journey times? Indeed does it actually have this data?
This brings me back to the one thing I did differently, I turned on ‘display speed limits’. What I found was not good. For one third of my journey CoPilot didn’t appear to know what the speed limits were, this section was entirely on secondary roads. When I got to the primary routes it didn’t get much better, it displayed the speed limits but they were nearly always wrong. One bit of this was expected (Thame to Wheatley) because the speed limit only changed about 6 months ago. The section from Oxford to Woodstock is more worrying – the 50mph speed limit has been in place for at least 10 years and CoPilot was reporting it as 60mph.
So the conclusion appears to be that CoPilot is using inaccurate data (both road and speed) to calculate journey times. Hardly surprising then that it comes up with inaccurate results.
I don’t know if NAVTEQ can supply the speed limits on secondary routes and Alk simply didn’t purchase this data or if NAVTEQ can’t supply it at all. My guess is that they can supply it, but Alk chose not to purchase it. If that is the case I can see some logic behind that decision, Alk is an American company and having driven in the US quite a bit my experience is that driving from almost anywhere to almost anywhere else you are nearly always using primary routes. Bearing in mind calculation speed, map file size, and the cost of the final product it therefore makes sense not to add speed limit information for secondary routes. Driving in Europe is very different though, for just about every journey I take a very significant part is on secondary routes.
David, I have (at your request) tried to supply some detail about the problems I am experiencing. If you could make the same effort to formulate a detailed reply, including things like what exactly is expected in the next update it would be much appreciated. And not just by me I think.
As a last note I believe NAVTEQ supply the maps for Garmin. Most people I have spoken to in the UK think that TomTom has more accurate maps.