Smoothing of trajectories

[chraibi]

In many papers, we see that there is a need to pre-process the trajectories by applying smoothing filters like for example Savitzky-Golay filter.

Maybe we could add a class of filters to preprocess the trajectories.

Figure taking from Smoothing Trajectories of People’s Heads

[schroedtert]

Great idea! Do you have any particular filters in mind?

[chraibi]

I thought at the Savitzky-Golay filter,
the Central Moving Average (CMA) and the Spline Interpolation of Inflection Points (SIP).

[MaikFZJ]

See also discussion #301

Distinguish between

• removing jitter
• detecting steps
• calculate main movement direction by
  • using step detection
  • spline interpolation, inflection point (Bernhard Steffen) (1)
  • Savitzky-Golay filter (Jiayue Wang) (2)
  • moving convex hull (Jana Pick, Janine Klein) (3)
  • moving average

One should take velocity into account when looking for step/stride.

(1)
https://arxiv.org/abs/0911.2165
subfolder in old JPSreport

(2)
Wang J, Weng W, Boltes M, Zhang J, Tordeux A and Ziemer V 2018 Step
styles of pedestrians at different densities J. Stat. Mech. Theory Exp. 2018
023406
Wang J, Boltes M, Seyfried A, Zhang J, Ziemer V and Weng W 2018 Linking
pedestrian flow characteristics with stepping locomotion Phys. A Stat. Mech. its
Appl. 500 106–20

(3)
https://juser.fz-juelich.de/record/888120

Some source about smoothing can be found in
https://juser.fz-juelich.de/record/888120

[MaikFZJ]

Have also a look to:
Jia, X.; Feliciani, C.; Yanagisawa, D.; Nishinari, K. Experimental Study on the Evading Behavior of Individual Pedestrians When Confronting with an Obstacle in a Corridor. Physica A: Statistical Mechanics and its Applications 2019, 531, 121735. https://doi.org/10.1016/j.physa.2019.121735.