Hi! We see you’re using an ad-blocker. We’re fine with that and won’t stop you visiting the site.
But as we’re losing ad-revenue from this then why not make a donation towards website running costs?. Or you could disable your ad-blocker for this site. We think you’ll find our adverts are not overbearing!
Joined: Aug 25, 2004 Posts: 634 Location: Lincolnshire, England
Posted: Sat Apr 23, 2005 11:33 pm Post subject:
MC’s “RoadWeightAdjust” settings improve only those journeys that need to exclude and include the particular types of road that his settings include and exclude. It follows that they will not improve journeys that require different road types to his settings; indeed, they may well worsen them.
What is still needed is for CP 5 to produce shortest, or quickest, journeys regardless of road type. This simply cannot be achieved with CP 5 as there is no combination of Road Type Bias settings that will do it – at least on the Pocket PC, and probably also on other platforms as far as I can tell from this thread.
Quite what use anyone would ever have for these settings seems to be anyone’s guess in any case. The supposed facility to, for example, 30% avoid motorways, 40% favour “secondary roads” 20% favour “local roads”, etc is born of a comprehensive misunderstanding. To illustrate the point, please think of a route and ask yourselves, where should the route go if I set 30% avoid one type of road and 40% favour another? Navigation software cannot be based upon the same logic as cooking recipes. Calling for 30% water, 40% meat, 20% vegetables and 10% other ingredients might well make sense when making a casserole, but the concept is completely useless and meaningless when trying to describe travel routes.
For overall set-up, each road type needs to be either included or avoided – completely. The application should then find the shortest or quickest journey based upon this combination of complete inclusions or exclusions. Waypoints, not system settings, are (or should be) there to set any detours from the route thus found.
Sadly, therefore, much as I wish this were a feature rather than a bug, it is far from it.
Perhaps it would be best to deal with concerns about ALK’s and others’ approaches to these issues on a separate thread as far as possible. Therefore, please see the new thread entitled COPILOT 5: AFTER-SALES SERVICE, USER ADVICE, etc.
Joined: Nov 24, 2003 Posts: 1441 Location: Swansea
Posted: Sun Apr 24, 2005 12:46 pm Post subject:
Borrowing from suggestions on this and another thread, I have arrived at a set of weightings and speeds which seem to conform reasonably well with typical road conditions. I have inserted these as appropriate in DEFTRIP.DAT, usertrip.dat and GPSTrip.trp I don't any longer have quickest routes taking longer than shortest. Every combination of parameters is only a compromise - what we are trying to do is make a machine which is incapable of thought come up, by a set of mathematical manipulations, with a solution which conforms with our feeling about what is the best answer. I know from my own programming experience, more years ago than I care to admit, that a computer will only do what you tell it to do. Anyway, for what it's worth, here are the values I am using. For those of you who feel confident enough to try different values it might be worth a try (always retaining the option to reverse the changes by renaming the original files .old and putting them back if you don't like the mod
Joined: Aug 25, 2004 Posts: 634 Location: Lincolnshire, England
Posted: Tue Apr 26, 2005 9:18 am Post subject:
It is of course helpful if users post what might enable others to cope better with a defective application. However, there is a danger that ALK will continue to do nothing if it appears that users believe that there are satisfactory means of adjusting CP 5, or that the shortcomings in CP 5 are only to be expected from this genre of application.
This is the incorrectv impression left by Dave Burrows - formerly of this forum and now of ALK!!?? Unfortunately, the same impression is now being promoted by posts such as the one above, from Robin2.
By all means, fellow users, please do post your suggestions, but please also remember that Copilot needs to be amended, if not for you then for others.
It is inclear what Robin2 means by "conform reasonably" well or "typical road conditions". However, tweaking road biases WILL NOT suit anything other than specific road type preferences, and it WILL NOT produce a universal solution or anywhere near it. While it is true that any computer programme works by "a set of mathematical manipulations", the mathematical logic employed does have to be sound for it to be of use.
Therefore, we need realise that once we exclude a road type from one journey, it will be excluded from all others until we "tweak again". Also, we need to beware of confusion between road speeds and road type biases. While going faster will reduce the time a journey takes (other matters remaining unchanged), it will not alter the distance travelled.
Joined: 24/06/2003 00:22:12 Posts: 2946 Location: Escaped to the Antipodies! 36.83°S 174.75°E
Posted: Tue Apr 26, 2005 10:58 am Post subject:
Robin2 wrote:
what we are trying to do is make a machine which is incapable of thought come up, by a set of mathematical manipulations, with a solution which conforms with our feeling about what is the best answer. I know from my own programming experience, more years ago than I care to admit, that a computer will only do what you tell it to do.
No, what we are trying to do is to get a machine to do some simple maths and come up with a reasonable approximation. Of course the definition of what is "reasonable" may be open to debate but I doubt that anyone would consider the routes listed below as reasonable.
The only thing we are telling the computer to do is to find the Quickest route between two points. The routing algorithm in the navigation software is expected to do the rest. Other navigation systems can deal with it, so why can't CoPilot?
I have tried so many settings and none of them have worked. This setting still fails 3 out of 4 of the (infamous?) "Skippy" tests too.
Quickest Route London to Dublin (routes via Belfast)
Quickest Route Southampton to Hastings (routes 182 miles M25 ANTI Clockwise via Potters Bar!)
Quickest M40 Junction 9 to M3 Junction 9 (30 Miles via A34 CORRECT!)
Shortest Knowstone to Chawleigh in Devon. (Gives 26 miles when the shortest is about 13)
Can other people try these routes and see what they get. _________________ Gone fishing!
It does better on the Southampton-Hastings, London-Dublin and Chawleigh to Knowstone routes (though in the latter it doesn't use the very tiniest roads). I am sure there will sometimes be inconsistencies in particular routes when the shortest may turn out to be longer than the quickest. The problem is that I am tweaking some parameters without knowing how their routing algorithm works, what weighting they give for right turns across the traffic flow etc. I have tried the above for quite a number of routes in and around my home town (Swansea) and it has invariably produced a sensible route. Not always the one I would have chosen, but I know which roads are full of sleeping policemen, which traffic lights give priority to traffic in the other direction, etc.
Suffice to say that after a week of using these values the routes are as sensible as those produced by TomTom (which I prefer, but which at present has less recent maps)
Robin
Joined: Aug 25, 2004 Posts: 634 Location: Lincolnshire, England
Posted: Wed Apr 27, 2005 12:02 am Post subject:
Robin2, I have run your suggested settings on some of my rural journeys that need to use "local roads" for Shortest routing. With or without your settings, CP 5 produces routes that exceed the necessary distance by 60% or more in most of these cases.
Could I ask you about the rationale behind your adjustments? I see that you have changed most, if not all, the urban road speeds. You have also set 20% "avoid motorways"; otherwise, your road-type biases are at their factory defaults.
You appear to believe that these adjustments should cause CP 5 to give
truly shortest routes for most situations. Is that what you meant and, if so, could you say how it could be that 20% avoiding motorways would correct routing errors where there are no motorways? Also, did you expect that the road speed adjustments would correct errors in Shortest routing, and, if so by what logic?
Could I mention to anyone who is not aware that it is not necessary to get into the car to test CP 5 and its settings, at least if you have the Pocket PC version of CP 5. That version, at least, has a Planning mode that enables routes to be tested in your armchair.
Joined: Aug 25, 2004 Posts: 634 Location: Lincolnshire, England
Posted: Wed Apr 27, 2005 12:16 am Post subject:
Sorry, but I should have added that the Pocket PC version is accompanied by a Desktop PC version, so you can even test your routes on the "big screen". I know this may be obvious to many, but users have indicated that they are testing their settings while driving in their own area, and taking a week to do it. Using these planning modes means that dozens of routes anywhere in the UK (and beyond) can be tested in a few minutes without moving from the house.
This means that no one, whether a novice, a forum team expert or a member of ALK's staff, need be in doubt about whether a "tweak" is universally applicable before posting it on the forum.
No, what we are trying to do is to get a machine to do some simple maths and come up with a reasonable approximation. Of course the definition of what is "reasonable" may be open to debate but I doubt that anyone would consider the routes listed below as reasonable.
Unfortunately the problem of vehicle routing is far from simple mathematics. Typing "vehicle routing algorithm" into Google will give you an idea of how much intellectual effort is poured into this classic mathematical problem. An exact answer is possible using a brute force method but running on a PocketPC could take hours or days. Instead, to get an answer in a reasonable time most algorithms employ "heuristics" (rules of thumb). An example might be "motorways are quicker than other roads". It may not give the absolute right answer but it gives one in a reasonable time - that is the "feature". The algorithm is tuned for a particular platform to give a result in a reasonable time (probably by restricting the number of optimising iterations). I doubt there are bugs in the algorithm itself so it is unlikely that ALK will fiddle with it. They may in time try to make the algorithm more efficient but that is a major rewrite. This would allow more iterations in the same time leading to more accurate results.
PONDEROUS asked earlier what it means to 30% avoid one road type or 20% favour another. Each segment of a road (between two junctions) has a distance which is used in "shortest" calculations and a road type which is used to determine speed which gives a travel time (distance divided by speed) - this is used for "quickest" calculations. Using 20% avoid causes the computer to do its calculations using a distance or travel time increased by that amount. Note that these tweaks are standard functionality on the desktop. Hopefully they will become available on the PocketPC too.
Small caution when using the desktop to test road bias on different routes. As the desktop has more compute power ALK has (AFAIK) altered the optimising settings so it does not give identical results to the PocketPC. _________________ --Ken
HP iPAQ h2210 ROM 1.10.07 ENG
Windows CE 4.20
Sandisk Ultra II SD 512Mb
Haicom 303MMF and BT-401 slipper
Copilot Live 6.0.0.68
Joined: Aug 25, 2004 Posts: 634 Location: Lincolnshire, England
Posted: Wed Apr 27, 2005 6:49 am Post subject:
KenS, with respect, what you say does not take us forward. The idea that has been floating around is that by tweaking these parameters, CP 5 will some how be caused to produce shortest or quickest routes. "30% avoid" etc. plainly will not achieve this; furthermore, what sort of route one can expect from it is a complete guess.
If road speeds are used in calculating Shortest routes in CP 5, then that is complete nonsense. It is like saying that if we put our foot down, we will cover a shorter distance - or get there before the fuel runs out!!.
The only maths needed in Shortest routes calculations, except in avoiding blocked road, one way streets etc., is DISTANCE and simple addition.
Having the Desktop PC tweaks on the Pocket PC will achieve nothing but to perpetuate the confusion.
Joined: Aug 25, 2004 Posts: 634 Location: Lincolnshire, England
Posted: Wed Apr 27, 2005 7:16 am Post subject:
Sorry, KenS, I now see that you were not suggesting that road speeds were used in Shortest routing.
It remains, however, that your explanation concerning percentage adjustments to "avoid" and "favour" does not stand up. Neither does the idea that the "tweaking facility" presently found on the desktop version will provide the solutions. By the way, if it is capable of doing so, then where are the solutions? We have had several combinations of settings posted during the past few days, but none of them works.
It is perhaps appropriate to pose the simple question, once more: what is the combination of settings that will universally cause CP 5 to produce Shortest and Quickest routes, or as near as makes no difference? I would be delighted to do a parachute jump for a charity nominated by the person who can produce it.
The assumption that something works is insufficient grounds upon which to recommend it - surely?
Joined: 24/06/2003 00:22:12 Posts: 2946 Location: Escaped to the Antipodies! 36.83°S 174.75°E
Posted: Wed Apr 27, 2005 8:24 am Post subject:
KenS wrote:
An exact answer is possible using a brute force method but running on a PocketPC could take hours or days.
Hmm, I have an old Garmin GPS-V which has about 1% of the CPU power of my SmartPhone and it still manages to produce reasonable routes. Other vendors (TomTom, Mapopolis etc) manage to produce reasonable routes on the Smartphone.
KenS wrote:
I doubt there are bugs in the algorithm itself so it is unlikely that ALK will fiddle with it.
It isn't faulty map data and it isn't something the user is doing, what do you think the problem is then?
I think the routing algorithm in CoPilot is poorly implemented, produces bizzare routes, ignores certain road types and it uses too much memory causing the application to crash. All of these assertations have been well documented in this forum and can be reproduced.
I have tried lots of things but I have not seen ANY tweak which will make CoPilot give reasonable routes. I would love to see a fix for this, but before anyone else posts some more random numbers for RoadWeightAdjus and DefRoadSpeeds can you please try the Skippy Tests listed earler in this thread. (No offence intended to the people who are spending a lot of time working with us to fix this problem, but I have tried so many things and nothing has worked). _________________ Gone fishing!
Joined: Nov 24, 2003 Posts: 1441 Location: Swansea
Posted: Wed Apr 27, 2005 1:29 pm Post subject:
With respect, the values I chose in DefRoadSpeeds were not random - I selected a range of speeds which I hoped would reduce the inbuilt preference for motorways which resulted in, for example, the crazy route from Southampton to Hastings. In the "Skippy Tests" that appears to have been achieved (though there is still a reluctance to use the very minor roads in the Devon example). I am quite sure that it would be possible to find other journeys where the routes would be less than ideal (as is sometimes the case with TomTom)
In real life tests in South Wales and southern England the parameters I quoted on April 27th have given me sensible routes, nearly always the same as TomTom produces. If they don't work for you, I have no problem with that, but I'm going to leave well alone on my own PDA
Robin
Joined: 24/06/2003 00:22:12 Posts: 2946 Location: Escaped to the Antipodies! 36.83°S 174.75°E
Posted: Wed Apr 27, 2005 4:06 pm Post subject:
Robin2 wrote:
In real life tests in South Wales and southern England the parameters I quoted on April 27th have given me sensible routes, nearly always the same as TomTom produces. If they don't work for you, I have no problem with that, but I'm going to leave well alone on my own PDA
OK, I'll have a go at re-installing CoPilot and then using your settings to see if I can get it to work. I am certainly not ruling out the possibility that there may be something wrong with my install. 8O
Sorry if I sound grumpy, it's just that I have spent a lot of time messing around with this thing trying to get it to work! _________________ Gone fishing!
Joined: Nov 24, 2003 Posts: 1441 Location: Swansea
Posted: Wed Apr 27, 2005 6:58 pm Post subject:
OKSkippy - I can understand why you are grumpy!! The ones I am currently using were quoted on the post which says 27 April (which I actually posted on 26th) They are
It is reluctant to use the very smallest roads, either for shortest or quickest. That's not altogether a bad thing though - TomTom has sometimes tried to send me on some awful cart tracks, sometimes unmetalled. I was in France earlier this month and TomTom pointed me down a gravel track with a wooden bridge with a 1 tonne weight limit! Admittedly it saved 2 miles, but I drew the line at following the instructions.
Posted: Today Post subject: Pocket GPS Advertising
We see you’re using an ad-blocker. We’re fine with that and won’t stop you visiting the site.
Have you considered making a donation towards website running costs?. Or you could disable your ad-blocker for this site. We think you’ll find our adverts are not overbearing!
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum
Or you could disable your ad-blocker for this site. We think you’ll find our adverts are not overbearing!
Hi! We see you’re using an ad-blocker. We’re fine with that and won’t stop you visiting the site.
But as we’re losing ad-revenue from this then why not make a donation towards website running costs?. Or you could disable your ad-blocker for this site. We think you’ll find our adverts are not overbearing!