cyang-kth/fmm

Strange Matching Results with Circles or Multiple Routes

Haicang opened this issue · 0 comments

Thanks so much for your work for map matching. However, I get the following strange results.

Describe the bug
I get strange matching results with circles or multiple routes. The problem I encounter can be visualized in the follow figure (matched trajectory, matched points, matched route), where the red lines (points) represents the raw trajectory and the blue line (points) represents the matched trajectory.

matched trajectory, from `pgeom'
matched_traj

matched points, from pgeom
mathed_points

matched route, from mgeom
matched_route

A similar issue can be found in #76 . However, I think the trajectory in my data do not have large noise.

Expected behavior
A matched trajectory without additional circle.

Program setting
I use latitude and longitude as the unit.
For ubodt:

<parameters>
  <delta>0.2</delta>
</parameters>

For fmm:

  <parameters>
    <k>8</k>
    <r>0.003</r>
    <pf>0</pf>
    <gps_error>0.0005</gps_error>
  </parameters>

When I switch to stmatch, I get the same results. I also tried pytrack, another map matching tool, but get the same results.