spline_interp {animalEKF}

Bezier spline interpolation of observations.

Description

Calculate a Bezier spline interpolation of irregular observations to regular-length time intervals.

Usage

spline_interp(di, area_map=NULL, t_reg=NULL, reg_dt=120,
              max_dt_wo_obs=60*30, maxStep=NULL, 
              centroids=matrix(c(0,0), ncol=2),
              nstates=2, spline_deg=3, split_logv=-3)

Arguments

di

Object of class data.frame containing irregular-spaced observations. Dataset must contain the following fields: 1) "X" and "Y": X and Y location coordinates, 2) "date_as_sec": time/date of observation as seconds since an epoch. "time_to_next" should also be included, but will be calculated if not. Also "tag" (animal identifier) should be included as well. If not, all observations are assumed to be of the same animal. This can be generated from sim_trajectory_joint with gen_irreg=TRUE.

area_map

Shapefile that all interpolated points should be inside of.

t_reg

Desired time steps (must have a constant difference) to interpolate to. If is given, the default value of reg_dt is overridden. Will be truncated to the set of values within the range of observed values of di$date_as_sec.

reg_dt

Length in seconds of each regular interval.

max_dt_wo_obs

When interpolating, the maximum time length without observations for a given shark that we will interpolate. If this is exceeded, algorithm will wait until next observation and start from there.

maxStep

Maximum number of regular steps to interpolate.

centroids

Matrix with two columns specifying the centroids of regions. If NULL, only one region will be used.

nstates

Number of behavioral states. For now restricted to a maximum of 2.

spline_deg

Degree of spline. The default is 3, or a cubic. Every spline_deg+1 observations will be used to construct one spline segment.

split_logv

If nstates=2, state 1 from the interpolated values will be designated by the logvelocity being < split_logv.

Value

d

Array of regular step locations.

di

Original irregular-step dataset.

shark_names

Vector of the names of sharks in the dataset.

d_ds

Output regular-step dataset d in form data.frame.

Author(s)

Samuel Ackerman

References

Bezier R package. Aaron Olsen.

Examples

#can also be 'di' output of sim_trajectory_joint (set gen_irreg=TRUE)

di <- data.frame(X=runif(n=9), Y=runif(n=9),
                 time_to_next=c(2,4,15,8,5,18,3,5,NA))
di$date_as_sec <- c(0, cumsum(di$time_to_next[-9]))
region_centroids <- cbind(X=runif(2), Y=runif(2))

#one log observation with dt =18 > 16 will be omitted
spl <- spline_interp(di=di, area_map=NULL, reg_dt=3, max_dt_wo_obs=16, maxStep=NULL, 
                     centroids=region_centroids, nstates=2, spline_deg=3, split_logv=-3)

plot(di[,c("X","Y")], xlim=c(0,1), ylim=c(0,1), type="b", las=1,
     "Observations interpolated by regular interval spline")
lines(spl$d_ds[,c("X","Y")], type="l", col="red")
legend("topleft", col=1:2, legend=c("observations","spline"), lty=1)