Guess start of escape — guess_escape_begin

This function determines the time point when the player was first away from the reference position, and then determines the last time that its velocity away from the reference position was smaller than a threshold. Positions are initially median-filtered, and so is the computed velocity. All calculations are performed without resampling.

guess_escape_begin_time(
  df,
  ref_pos,
  esc_dist = 0.75,
  span = 5,
  min_speed = 0.1,
  min_time = min(df$time),
  max_time = max(df$time),
  indx = 1
)

Arguments

df: movement data frame
ref_pos: reference position, a list containing `pos_x` and `pos_z`
esc_dist: minimum distance from reference position to determine escape
span: span of the median filter
min_speed: minimum speed to determine start of escape
min_time: minimum time to start searching for escape
max_time: maximum time to search for escape
indx: currently disabled (only needed for debugging)

Details

Developer note: if escape is imposed on an ongoing movement away from bush, it might not be detected by this algorithm based on a fixed velocity criterion. However, it can actually be ambiguous to decide whether the escape was planned before or after the min_time. The algorithm could potentially be improved by using an acceleration criterion (e.g. zero crossing) but the tracker data are a bit noisy and this would require some more elaborate signal processing/filtering.