Forum: General Topics

ranchodoug said" About 2/3 of my trackpoints are within 10m of their true position as shown on google earth, close to 1/3 are within 20m, and a few percent can be up to 80m off. What's odd is the errors aren't random."

I have noticed this also, but I am NOT absolutely positive that the Google maps and indicated GPS positions are perfectly accurate and are the true positions..I know that the GPS positions shown by TMJ match the GPS positions shown on dedicated stand alone devices that my hiking partners have.

So is it a GPS error or is it a Google error????

.

MrTom, Stephen,

Thanks. Most errors I've seen appear to be about expected, but errors up to 80m seemed high. The extreme errors tended to be in the 'Who Cares?' regions of the map so perhaps there were no GPS/image correction control points in the vicinity.

My minimum trackpoint (TP) to TP distance has to stay small. I like to walk at the Wild Animal Park in San Diego for exercise. The paths twist/turn frequently so a 10m-15m minimum distance creates a terribly distorted linear approximation of the true track. TMJ is sort of like my pedometer, although it's quite interesting to retrace my steps. With a GPS update rate at once per sec and with a minimum track distance to 1-2m the track is excellent. The few random TP clouds when I stop are quite acceptable. The clouds are almost like waypoints.

That is, until last Sunday. Unfortunately the Activity Profile was set to Driving (default?) while I was walking (the Profile has such a large affect on the track that I think changes to the Profile should require confirmation). Even allowing for the wrong Profile, the tracks were at times extremely erratic and bad. If I believed the tracks, I was both meandering (off by up to 600m) and in brief spurts running (at isolated points over 200 km/hr) through the Asia and Africa paddocks of the Park. There are rhinos and cape buffaloes in those paddocks. I don't remember any large beasts chasing after me, so was likely not at the reported locations. Stephen, if you look at 'Auto-Track (Sun 18th Mar 2012)' you'll see what I mean. The large green/tan areas are the open areas where the herbivores (small and large) roam free. But, if the GPS positions were actually correct, points in the 176 to 185 region were well inside the lion pen and I also don't recall being chased by Lions, so probably wasn't there either. GPS errors were extreme that day.

It was drizzling, not really raining, but wet. The phone was inside my left breast shirt pocket, face to me, covered by a light jacket. Have you seen light rain affect GPS by such an extreme? The GPS errors were far more serious than I've seen in the few weeks I've been experimenting with TMJ. I didn't bother looking at the GPS logs since every time I've looked, TMJ was faithfully using every GPS position reported by the phone. As you mentioned, once the track was off by a large margin, it tended to stay off track with the same bias error for some time, then recovered.

The systematic errors I mentioned previously were after I'd been driving for at least 15 minutes. I don't think the car has an electric windshield. If so, it wasn't turned on. The phone is usually on the center console, face up, while driving. The ground I'm driving/walking on isn't typically that steep, but there are sometimes cliffs nearby so maybe there is some multipath. This last weekend the errors are particularly bad so maybe it was a combination effect.

If these errors are typical of a BBerry 9930 I'll just live with them, but wondering if they were experienced by others.

Thanks,

Doug

I wish we could post graphic attachments. I took a hike yesterday and had TMJ set to "walking" mode. Forgot to turn it off as we drove home. The track when reviewed in Google Hybrid was right on from the start point, through a wash bed back to the truck location. It was correctly shown on the return highway drive exactly to each of our houses.

I have an old Blackberry Pearl 8120 with an external Freedom2000 GPS. I am very satisfied with its accuracy.
At other forums I have seen numerous complains about the inaccuracy of internal phone GPS.
I tend to question the accuracy of the various maps and there might even be inaccuracies in the maps with respect to who made them and where they were made.

As indicated, yesterdays track was correctly shown in the Google Hybrid map. Yet when I take a walk in my local neighborhood the Google maps are about 15 feet to the East with respect to the GPS track. Seems inconsistent as yesterdays track took me exactly to my house.

I know the maps in my area are made by a company that drives around in a Hummer SUV with a GPS unit on the roof. They then transmit the GPS and graphical data to Google. Haven't been able to talk to them yet about what calibrations they make to the equipment.
Yesterdays track showed a max elevation of 1701.2 meter which = 5 581 feet. Delorem Topo USA show that point as 5989 feet and graphically puts us on top of a mesa that I know we did not reach.

Consumer GPS units, data, and maps are prone to errors( some intentional ) and we have to accept that fact for now. In the USA the data comes from government owned satellites. In China the maps are intentionally m1s-calibrated. Hopefully that will change with the advent of private GPS satellites.

Stephen,

I'm not sure if turning off Allow Costs and Location Assist, or just no rain solved the problem but a track this morning was just fine. I guess I'll just monitor the evolving track more carefully in the future.

Thanks for you help,

Doug

Hi Doug,

Mmm, thats interesting - as you say, its not definitively the Allow Costs thats done it (as well as the weather it could just be that the satellites were in worse positions initially - they are continously moving so at sometime there are more overhead than at other times). Be interested to hear how the monitoring goes, and maybe switch the Allow Costs/Location Assist back on at some point to see if the problems come recur...

Cheers,
Stephen