GitHub

API

Wahoo.track — Function

Tracks the location of the fish

track(;pos_init::Matrix,
       tsave::AbstractVector = 1:100,
       bathymetry::GeoArrays.GeoArray,
       observations::Vector,
       observation_models::Vector{Function},
       sensor_positions::Vector,
       spatial_resolution,
       movement_std,
       smoother::Bool = true,
       filter::Bool = false,
       n_trajectories::Int = 0,
       show_progressbar::Bool = !is_logging(stderr),
       precision = Float32)

Infers the location of the animal based on a diffusion model and smoothing.

Keyword Arguments

  • pos_init::Matrix: Initial probability distribution of the fish position.
  • tsave::AbstractVector: Time steps at which the probability map is saved.
  • bathymetry: Bathymetric data as GeoArray.
  • spatial_resolution: the spatial resolution [m] of the bathymetry.
  • movement_std: Standard deviation of the fish movement within one time step [m].
  • observations: Vector holding all observations. Each element contains the observation of a separate sensor.
  • observation_models::Vector{Function}: Vector containing the observation model for each sensor.
  • sensor_positions: Vector of tuples of coordinates or nothing, i.e. Vector{Union{Nothing, Tuple{Real, Real}}}.
  • smoother = true: if true the probabilities from the smoother run are returned.
  • filter = false: if true the probabilities from the filter run are returned.
  • n_trajectories = 0: Number of trajectories to sample.
  • show_progressbar = !is_logging(stderr): defaults to true for interactive use.
  • precision = Float32: numerical floating point type used for computations.

Note, the elements of the vectors observations, observation_models, and sensor_positions must be sorted in the same way, i.e. the elements at the same position in the Vectors refer to the same sensor.

Return

A named tuple with the following elements:

  • log_p: Log probability of the observations, $\log p(\mathbf{y}_{1:T})$.
  • tsave: Vector of time steps at which the results are saved.
  • trajectories: Sampled trajectories if n_trajectories > 0, otherwise nothing.

Additionally, if smoother = true:

  • pos_smoother: $\Pr(\mathbf{s}_t \mid \mathbf{y}_{1:T})$, the smoothed probability distribution of the fish positions for all timesteps in tsave.
  • residence_dist: Residence distribution, $\frac{1}{T}\sum_{t=1}^{T}\Pr(\mathbf{s}_t\mid \mathbf{y}_{1:T})$.

Additionally, if filter = true:

  • pos_filter: $\Pr(\mathbf{s}_t \mid \mathbf{y}_{1:t})$, the filtered probability distribution of the fish positions.
source