Linear Model Type #7
Conversation
…ab/RobotDynamics.jl into linear_model new changes from master
There was a problem hiding this comment.
I think this is really close. I'd like to clean up the linearization/discretization methods.
I think it may also be a good idea to have a simple AbstractLinearModel that both LinearizedModel and LinearModel inherit from, just so you can easily cover both with a single dispatch, since they are nearly identical in practice.
src/linearization.jl
Outdated
| @assert is_timevarying(model.linmodel) == true | ||
| @assert get_times(trajectory) == (model.linmodel.times) | ||
|
|
||
| if is_timevarying(model.linmodel) |
There was a problem hiding this comment.
This check is redundant given your previous assertion that it is time-varying
src/linearization.jl
Outdated
|
|
||
| Linearize `nonlinear_model` about `trajectory` and discretize with given integration type `Q`. Returns a [`LinearizedModel`](@ref) | ||
| """ | ||
| function linearize_and_discretize(::Type{Q}, nonlinear_model::AbstractModel, trajectory::AbstractTrajectory; is_affine=false) where {Q<:Explicit} |
There was a problem hiding this comment.
Can we simplify these methods to an interface similar to LinearModel? I think it would make sense to have a powerful LinearizedModel constructor that handles continuous/discrete, affine/non-affine, and then have an update_reference!(model, Z) that re-linearized the model about the updated reference trajectory Z (could be a single KnotPoint).
I agree that it could be nice to have an |
This PR adds the
LinearModeltype, as well as aLinearizedModeltype useful for linearizing a nonlinear model.