| Versions |
  |
|
| Author |
 Topic  |
|
anthlover
2307 Posts |
 Posted - 15 juin 2004 : 12:13:39
|
RE: Garmin, Magellan, TomTom, etc. how well does Dead Reckoning work??
Please share your experiences with Dead Reckoning capable GPS.
As of 06/15 we have:
Garmin SP2650 *
Garmin SP2660 *
TomTom GO **
* gyro + dealer installed speedo connection
** integrated solid state gyro/accelerometer |
Edited by - anthlover on 15 juin 2004 12:14:43
|
|
|
Ads
|
|
|
gpspassion
84988 Posts |
|
|
alextse
98 Posts |
Posted - 15 juin 2004 : 13:19:15
|
OK, aside from my recent Garmin 2650 purchase, my experience with other GPSRs are solely with Lexus OEM units. I was extremely impressed by the Lexus units, easy to read/use, announced directions were very clear, and I never ever noticed a signal loss (due to DR I'm assuming).
The Garmin 2650's DR is comparatively "pretty good". I don't even really notice when the unit goes into DR mode, the tracking acts just like when it has GPS lock. I've gone through the NYS Thruway, Lincoln Tunnel, downtown Manhattan, and about 5 continuous miles under elevated rail tracks with no guidance loss. However, there are a few times in "aggressive" city driving where I have managed to confuse the unit for a few seconds but it quickly corrects itself after about five seconds of straight driving. So it may not be perfect but that level of variance is quite acceptable at a $1300 price point.
As I've pointed out in other postings, I'm seriously considering a reradiating antenna so I can reduce my DR reliance even more. I guess I'm spoiled by the seemingly zero deviation I experienced on the Lexus GPS units.
I hope that was helpful. I'm more than happy to answer any specific questions regarding my DR experience. |
 |
|
|
gpspassion
84988 Posts |
Posted - 15 juin 2004 : 13:35:06
|
Thanks for the input, certainly looks quite impressive.
Couple of questions (and comments in the GO)
1. I take it the speed displayed goes up and down with the car (on the GO it depends, sometimes it works and sometimes it keeps going up!)
2. Have you noticed any "catching up" when you exit a tunnel for instance? (yes on the GO)
3. Does the display change when you are in DR mode? (yes on the GO - flashes ASN)
4. Is the difference with the Lexus maybe due to the fact that you're warned of DR where I would expect the Lexus system to "hide" that?
5. Does navigation pick up immediately when you exit a garage for istance? (no on the GO)
Not sure a reradiating antenna would help unless you find yourself losing the "GPS" fix pretty often? Holux have a neat one in their catalogue and they sent me a test model...no time to test it yet though :-(
___________________________________________
Discounts and Assistance/Réductions et Assistance (Club GpsPasSion)
Best "GPS" PocketPC/Choisir un PDA ou un Kit "GPS" / Où commencer? |
|
|
|
diverhank
USA
924 Posts |
Posted - 15 juin 2004 : 15:43:18
|
quote:
Originally posted by gpspassion
Thanks for the input, certainly looks quite impressive.
Couple of questions (and comments in the GO)
1. I take it the speed displayed goes up and down with the car (on the GO it depends, sometimes it works and sometimes it keeps going up!)
I expect this from the Tom Tom Go. I guess this depends on two things...the sensitivity of the accelerometer and the other is on how good the software is handling it. I suspect the problem is a little of both. It's tough to use a cheap accelerometer to do this sort of work. Let's say you are going along at a steady 60 mph (100km), the accelerometer should read nothing (zero) because there's no acceleration(no change in speed). Of course this accelerometer has been on for a while, bouncing around and what not. Unless there's a good algorithm to zero out the error, it is not likely to read zero. Now you lose GPS and the unit goes into DR, the accelerometer error (or bias) is enough to make the unit think you are changing speed. Also if the accelerometer is not sensitive enough and you change speed so ever slightly, it's not going to pick up the acceleration, causing errors.
I suspect that the Garmin does not have this problem because it reads the speed and the changes in speed from the speedometer.
Hank |
|
|
|
gpspassion
84988 Posts |
|
|
rwcx183
81 Posts |
Posted - 15 juin 2004 : 16:54:05
|
All integrated GPS/accelerometer/dead reckoning systems, continually null the bias in the accelerometer using GPS data, else the system performance would be so poor during GPS outage, that it wouldn't have been worth the cost to add the accelerometer in the first place.
J.G.
|
|
|
|
gpspassion
84988 Posts |
|
|
rwcx183
81 Posts |
Posted - 16 juin 2004 : 03:23:28
|
quote:
Originally posted by gpspassion
Right, but what happens when the GPS data is no longer available? I can live without an accelerometer when I have a GPS signal but could use one when I don't ;-) I guess there is a marginal use for an accelerometer simultaneaously with a GPS for better "cornering"?
___________________________________________
Discounts and Assistance/Réductions et Assistance (Club GpsPasSion)
Best "GPS" PocketPC/Choisir un PDA ou un Kit "GPS" / Où commencer?
Well, in the absence of GPS data, you're navigating purely on the accelerometer (gyro actually) and wheel ticks. This accuracy will gradually degrade over time, until you reacquire GPS position.
As for using the accelerometer and GPS data together at the same time, yes it is usually done by combining both information sources into a weighted composite solution. This is done through the magic of a Kalman filter, which would be designed in this case, to use each source of data with a greater or lesser weight, depending on the estimated error from each source. GPS data is quite good for slowly varying positions or headings, but it includes a significant amount of higher frequency noise that is not easily filtered, without incurring significant delay. On the other hand, the accelerometer data is exactly that, acceleration. By itself, not terribly useful, but if you mathematically integrate it over time, you get velocity, which is quite useful to the Kalman filter navigation solution. If you then integrate that velocity over time, you get position. That's also obviously useful. The integration over time, turns out to be quite a good filter, such that the higher frequency variations seen in the GPS data, are not seen in the accelerometer derived velocity and position. Such positions and velocities are not particularly accurate at low rates of change, such as when you're at a dead stop, or moving very slowly. In those conditions, you will see drift. Which is why we always use the GPS data to null out that drift, when we have GPS data. You can think of it this way, the GPS data is like a woofer loudspeaker and the accelerometer is the tweeter. Each is used to it's best advantage.
J.G.
|
|
|
|
alextse
98 Posts |
Posted - 16 juin 2004 : 03:46:59
|
quote:
Originally posted by gpspassion
Thanks for the input, certainly looks quite impressive.
Couple of questions (and comments in the GO)
1. I take it the speed displayed goes up and down with the car (on the GO it depends, sometimes it works and sometimes it keeps going up!)
2. Have you noticed any "catching up" when you exit a tunnel for instance? (yes on the GO)
3. Does the display change when you are in DR mode? (yes on the GO - flashes ASN)
4. Is the difference with the Lexus maybe due to the fact that you're warned of DR where I would expect the Lexus system to "hide" that?
5. Does navigation pick up immediately when you exit a garage for istance? (no on the GO)
Not sure a reradiating antenna would help unless you find yourself losing the "GPS" fix pretty often? Holux have a neat one in their catalogue and they sent me a test model...no time to test it yet though :-(
1) Yes, the speed display constantly changes. It seems quite accurate, more so than my cars speedometer. I've used it on long highways to set my cruise control and haven't been ticketed yet. BTW, with cruise control the GPS speedo is constant so that sounds right.
2) No problem in a tunnel or any other straight or gently curving situations. My only (slight) problem has been when I cross through four lanes of traffic and back again in NYC traffic in the space of a block...that tends to get the unit wondering as to my actual direction!
3) No display change unless 1) you change a tab to show satellite status or 2) you go into the GPS info sub menu. Otherwise the unit just keeps chugging along like nothing is wrong.
4) Oh, I'm certain that the Lexus GPSR hides the loss of signal info. However, I would say the Lexus unit is a lot more like the Magellan RM with data on a DVD. The units are tightly integrated with the car's other electronics (a plus) but otherwise the GPS is pretty but not very full featured. Not to mention it costs some $2500 to $3000 to have installed!
5) Hmm,good question, I haven't tried that.
Lastly, it's not that I find myself losing signal "a lot", but I just can't stand losing signal at all! That's why I got the 2650 in the first place. So if I can add a sub-$100 accessory that can improve reception dramatically (see anthlover's many posts on this) and is non-intrusive (versus having to constantly connect/disconnect) then I'll get it! |
|
|
|
anthlover
2307 Posts |
|
|
gpspassion
84988 Posts |
|
|
Mike Kiwi
United Kingdom
15 Posts |
Posted - 05 oct. 2004 : 15:20:29
|
Has anyone ever tried building a home made dead reckoning system using accelerometers like this one from Analog Devices? http://www.analog.com/en/prod/0,,764_800_ADXL203%2C00.html
You could intercept the NMEA data stream and augment the data with the dead reckoning when you lost gps reception. You would need accelerometers on 3 axis to get an accurate indication of the direction/speed but in theory it could be done.
I have been wondering about this for a while, but it seems that the preferred DR method is using a speed pulse so I wonder if accelerometers are workable. I now see that TomTom are using them in their "TomTom Go" system and would be interested to know what sort of distance you can travel with no GPS signal before the accuracy degrades to an unusable extent.
|
|
|
|
paulkbiba
USA
5024 Posts |
Posted - 05 oct. 2004 : 15:53:08
|
Unfortunately the system on the Go is not reliable. I have never had it work for long enough to take me over the lower level of the George Washington bridge or the Lincoln Tunnel (about 1.5 miles).
Moderator
Don't forget the GPSPassion Club! |
|
|
|
Mike Kiwi
United Kingdom
15 Posts |
Posted - 06 oct. 2004 : 01:55:13
|
quote:
Originally posted by paulkbiba
Unfortunately the system on the Go is not reliable. I have never had it work for long enough to take me over the lower level of the George Washington bridge or the Lincoln Tunnel (about 1.5 miles).
Roughly how far do you get before it goes completely off the tracks?
I am probably not that worried about longer distances, it's just when you lose Sat lock for 10 or 20 seconds between tall buildings etc and then miss your turn. |
|
|
|
gpspassion
84988 Posts |
Posted - 06 oct. 2004 : 02:02:16
|
Having used the Go extensively in Paris where tunnels abound, it's hard to determine a pattern in the way it behaves, but generally it takes a bit too long to kick in to be useful for quick lost fixes. Also they say it's a two place sensor but I've only really seen "acceleration" being reflected. These observations might come down to software settings...Most of the time speed is well reflected (acceleration/decelaration), but sometimes it zooms out of control!
___________________________________________
Discounts and Assistance/Réductions et Assistance (Club GpsPasSion)
Best "GPS" PocketPC/Choisir un PDA ou un Kit "GPS" / Où commencer? |
|
|
|
Topic |
|
|
|
| This page was generated in 0,62 seconds. |
 |
|
- 
del.icio.us
Dead Reckoning, how well does it work ?
