Fuzzy via Point

[krzysiek mbis]

Hi,
I'm testing fuzzy linking and I have one question. Consider the route near Kolkwitz

Now lets put fuzzy via point (with 2100 metres radius) in the middle of this route, and run route request again. We will se as below:

If we look a little bit closer we will see that route didn't take radius into consideration:

I thing that this route supposed to be the same with such a big radius.
In other routs (very similar to that one) it seems to work fine.
In fuzzy.zip I send you a request in detail. Could you be so kind and tell me how to do it right ?

Best Regards,
Krzysiek.

[Bernd Welter]

Hello Krzysiek,

I think I can explain the cause of the result: just look at the following internal workflow in case of

Code: Select all

<WaypointDesc linkType="FUZZY_LINKING" fuzzyRadius="2100" xmlns="http://xroute.xserver.ptvag.com">

Based on your NEXT_SEGMENT - FUZZY_LINKING - NEXT_SEGMENT with the waypoints[green / blue / red]:

• we compute the route from [GREEN] to [BLUE]
• we go back 2100m in reverse direction from BLUE, the result coordinate is BLUE**
• we compute the route from [BLUE**] to [RED]
• we concatenate [GREEN]-[BLUE**] with [BLUE**] to [RED]

Now let's consider the exact location of BLUE: it is not located on the direct path between GREEN and RED but on the opposite driving direction. The attached screenshots show the route from [GREEN] to [BLUE]

route from GREEN to BLUE performing the nasty manoeuvre at the highway exit.

Close visualization of the route - you can see that the coordinate of BLUE is located on the wrong side.

Usually this works fine in larger areas with walking in reverse on the same highway direction. In your sample the internal logic produces such a result.

If BLUE is defined by a click in a map the probability of "choosing the proper side of the highway" is 50:50.

Best regards Bernd