Forum: General Topics
Forums / General Topics / track mileage on gps errors
| Subject: | track mileage on gps errors | |
wojsyl 11:13 | Sometimes, when there's poor reception, upon entering a tunnel/building etc. the gps gets confused and sets you several metres or several hundred metres away in a split of second, then when the reception improves it gets back to the correct spot. This causes certain "peeks" in the tracks that I'm sure we have all experienced at times. This of course also influcences the track mileage, which at times can be annoying esp. when walking and the mileage suddenly increases by 1 km or such. Do you think TMJ-mobile could be able to automatically detect such situation (like sudden "teleportation" or unreasonable speed increas) and maybe start a new track section (and give an option to remove it) for this ? | |
ChrisM 15:15 Location: Phone Model: | Glad it's not just me that experiences this. Like you say, it can be quite irritating if your distance walked suddenly jumps up by 1/2 mile or something. It is possible to manually remove points in the graph screen (key 9), but I suppose it should be possible for TMJ to spot and mark somehow points that are a long way off the expected track, or that don't match the average speeds either side. ie If your track points are roughly say 50m apart, then there are a couple that are 200m apart then back to 50m, or if the speed on a couple of points suddenly jumps from 4mph to 12mph then back again... Over to you then Stephen... :-) | |
sentinal 16:36 Location: Phone Model: | I have use TMJ frequently when walking in the Cotswolds (lots of hills, tree cover). I get spikes with speed and altitude which make my tracks slightly inaccurate (not TMJ, poor nmea gps signal data). To prevent this, could an optional limit amount be entered before the activity you know your doing. For walking say 4 miles per hour, any speeds over would be taken out of the track recording. This will also help the accuracy of the Calories burnt which Stephen is working on. New track selection would be a good solution also. | |
Stephen 11:58 Location: Phone Model: | Hi guys, As ever, thanks for the positive comments/ideas. The GPS signal 'spikes' are something that I find annoying too, though I've yet to work out a reliable way of filtering these out. Although inputting a 'maximum track speed' might possibly aid in this, I'd rather not go down this route as it could lead to a situation where, for example, the user forgets to change this setting and ends up with not recording anything in the track at all. Having said this, there is a 'moving average' speed which TMJ stores internally, that essentially gives a very highly-smoothed average of the speed over the last few minutes. I've been thinking of using this value as some sort of 'threshold' for TMJ to decide whether it should add a new trackpoint or not. eg: if you're walking along at a fairly constant pace, this average will settle down to say 4mph. If say a new trackpoint would then result in a speed 'spike' of twice this (ie 8mph), then it would initially hold back from adding the point. However if the current speed then stays at around 8mph, the internal moving average would slowly rise to this value, and the next trackpoint request would be accepted. There might be side effects from this though, eg if you change from driving to walking, it would take a few minutes for TMJ to adjust to the new speed, during which time the filtering wouldn't work so well. (Maybe if manually starting a new section also resets the average speed this could help). Will experiment further...! Cheers, Stephen | |
sentinal 21:50 Location: Phone Model: | The moving average you have already incorporated sound great. David | |
sentinal 11:27 Location: Phone Model: | After recording many tracks through woods and valley's. I'm very impressed how accurately TMJ records altitude. | |
(You must be logged in to post a reply to this thread)