OpenStreetMap

Another shot at GPS navigation with OsmAnd+

Posted by maxolasersquad on 29 November 2012 in English (English)

This Thanksgiving my family and I made the trip up to North Carolina to visit an important friend.

We left Tuesday afternoon after I got off work at 5pm. Not wanting to "chance it" I used Google Navigation to get to some relatives in North Georgia where we spent the night. The next day I downloaded the offline data for Florida, Georgia, South Carolina, and North Carolina and plugged our destination in to OsmAnd and showed the directions to the people we where staying with and they agreed that the directions where accurate.

For this part of our trip the navigation worked very well. It took us to the cabin rental office, and from there, to our cabin, without problem. While in our mountain cabin I took some time to correct some of the mapping in the area. I rode was named incorrectly (Google had the same incorrect name) and I added additional detail.

When our visit finished, I asked OsmAnd to take us to Atlanta, where my wife was to drop me off so I could attend a concert before taking the bus home. At this point things did not go so well. It spent a little time trying to calculate our route, got to 100% and then just sat there. After growing a little impatient I canceled and retried and it successfully calculated the route the second time. This is where things became really problematic. Even though I was following the suggested route, it kept recalculating on me. Since calculating a route that long offline takes a little over a minute, this became stressful as I had to make sure I knew how far my next turn was, and which direction to turn.

At one point it started telling me, in order to make a left at an intersection that was only a couple of miles ahead, I should instead turn around and make a ~50 mile detour to some completely different interstate. Of course I ignored it and made the proper turn, but it still wanted me to turn around, go back from where I came and then enter the intersection again to take me back to where I already was. At that point I manually cleared the directions and told it to recalculate again, which it did, and then started to recommend a completely different, longer, route to Atlanta.

At that point I put Google Maps back on, which directed me using the same route OsmAnd originally told me to take.

When OsmAnd was acting right it was very cool. At one point I was even able to get directions when my friend couldn't because there was no cell signal, and thus, no Google Maps. In this way OsmAnd was very helpful as we where in the middle of nowhere and was not sure which way to go.

However, I still don't feel that OsmAnd is quite there yet in usability, stability, or reliability to be recommendable over the Google Maps application. I am not including in this recommendation the state of the OSM data, which is out of the hands of the application itself.

However, I am very pleased with how far along it has come since the last time I tried this experiment over a long distance trip. At this rate OsmAnd is shaping up very rapidly. I do use it in day-to-day travel as I zip around town, which is very helpful in finding errors in the map data itself.

Location: Capital Circle Office Center, Tallahassee, Leon County, Florida, 32399, United States of America

Comment from Sanderd17 on 29 November 2012 at 19:17

The problem is that OsmAnd doesn't have servers to rely on. All calculation has to happen on the device. Unless you use an online routing service. In that case, you will be able to calculate longer routes.

The "turn back" issue is also known. To save calculation, a part of the original route is kept. So if you deviate from that route, OsmAnd will try to find a new route towards that route, and not directly towards the end point.

But long distance trips will be a problem until phones get a lot faster I think.

Hide this comment

Comment from maxolasersquad on 29 November 2012 at 19:50

After the route is calculated, and as long as I follow that route, there is no reason for it to ever recalculate, regardless of the processing power of my phone. If I do deviate from that route, and I ever make it back on to that route, it should be able to realize that no tracing back is needed. It could simply discard the part of the route that is "behind" me. This would take some processing, but nothing significant.

I didn't mind waiting around a minute for it to figure out the calculations initially. I would expect even a modern desktop PC to need some time to look at all of that data and figure out a good route. But once a route is calculated, and I stick to that route, there's no technical reason for any problems.

One possible issue that may have confused OsmAnd could be road alignment. If these old country roads I was on are not aligned properly, based on sloppy unreviewed Tiger data, I could see the phone thinking I was off the road, and then think it needs to recalculate that route since it would look like I had in fact strayed from the calculated path.

Hide this comment

Comment from aseerel4c26 on 30 November 2012 at 00:47

Thanks for the report, interesting (despite I own no Android device [yet])!

Hide this comment

Comment from porjo on 3 December 2012 at 21:34

I use Navfree on my phone (which also uses OSM data) and it routes and re-routes just fine without requiring any Internet connectivity. The OSM data is pre-compiled into Navfree's own (proprietary) format and I imagine would be optimized for routing. Perhaps OsmAnd's data needs more optimisation?

Hide this comment

Leave a comment

Parsed with Markdown

  • Headings

    # Heading
    ## Subheading

  • Unordered list

    * First item
    * Second item

  • Ordered list

    1. First item
    2. Second item

  • Link

    [Text](URL)
  • Image

    ![Alt text](URL)

Login to leave a comment