added: new KalmanCovariance struct to handle covariance sparsity efficiently

[franckgaga]

Similarly to #202 for the objective function weights in MPCs, a new KalmanCovariance parametric struct is created. First, it avoid duplicate code in the constructor of the Kalman filters. Second, two new parameters are introduced to preserve special types like Diagonal{NT, Vector{NT}}. It will also preserves other special types like SparseMatrixCSC. The performance advantages will be mainly visible for the MovingHorizontEstimator with nonlinear plant models, since the objective is computed with e.g. dot(V̂, invR̂_Nk, V̂) and the operation is faster when invR̂_Nk is a Diagonal matrix.

[codecov-commenter]

Codecov Report

Attention: Patch coverage is 95.27027% with 7 lines in your changes missing coverage. Please review.

Project coverage is 98.61%. Comparing base (336eaf3) to head (73e36ce).

Files with missing lines	Patch %	Lines
src/estimator/construct.jl	83.72%	7 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #216      +/-   ##
==========================================
- Coverage   98.70%   98.61%   -0.09%     
==========================================
  Files          26       26              
  Lines        4308     4327      +19     
==========================================
+ Hits         4252     4267      +15     
- Misses         56       60       +4     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.