Change names of lqr -> lqrc, dlqr -> lqrd, care/dare -> arec/ared

[baggepinnen]

The function lqr still exists for systems as inputs, where it's clear from the system type which version to call. For matrix inputs, we now require an explicit c/d.
The c/d has been put at the end of the function name instead of in the beginning to be consistent with MatrixEquations and to aid discovery.

lyap still exists without a c in the end since the function belongs to LinearAlgebra.

[JuliaControlBot]

This is an automated message.
Plots were compared to references. 11/11 images have changed, see differences below.
After pulling this PR, please update the reference images by creating a PR to ControlExamplePlots.jl here.

Difference	Reference Image	New Image
✔️ 0.006
❌ 0.048
⚠️ 0.016
✔️ 0.0
✔️ 0.002
✔️ 0.002
✔️ 0.003
✔️ 0.004
✔️ 0.012
✔️ 0.0
✔️ 0.011

[codecov]

Codecov Report

Merging #563 (1b0b60c) into dev (f644db6) will decrease coverage by 0.10%.
The diff coverage is 86.36%.

@@            Coverage Diff             @@
##              dev     #563      +/-   ##
==========================================
- Coverage   85.08%   84.98%   -0.11%     
==========================================
  Files          31       31              
  Lines        3192     3184       -8     
==========================================
- Hits         2716     2706      -10     
- Misses        476      478       +2     

Impacted Files	Coverage Δ
src/ControlSystems.jl	100.00% <ø> (ø)
src/timeresp.jl	89.51% <ø> (-2.16%)	⬇️
src/matrix_comps.jl	88.01% <83.33%> (ø)
src/synthesis.jl	88.46% <83.33%> (ø)
src/simplification.jl	100.00% <100.00%> (ø)
src/types/lqg.jl	83.80% <100.00%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update a0c960a...1b0b60c. Read the comment docs.

[imciner2]

Changing dlqr to lqrd will really mess people up who are coming from Matlab. Those two functions in Matlab are very different - dlqr designs the gain assuming a discrete-time system and a discrete-time cost function whereas lqrd designs the gain assuming a continuos-time system and continuous-time cost function (just with both discretized internally, so the discrete-time controller has a cost equivalent to the continuous-time cost). The current implementation matches Matlab's dlqr function, not its lqrd function.

I'm not really a fan of the other changes either, I kind of liked the prefix for differentiating between the continuous and discrete-time versions.