Add blog post: Centimeter-Level Positioning with GTSAM: Double-Difference Factors for RTK GNSS

[inuex35]

This PR adds a blog post introducing the double-difference GNSS factors from borglab/gtsam#2502.
The post covers:
Why double differencing cancels satellite/receiver clock biases and atmospheric terms
The four new factors (DoubleDifferencePseudorangeFactor, DoubleDifferenceCarrierPhaseFactor, and their lever-arm variants)
A factor graph diagram for tightly-coupled GNSS-IMU positioning

[dellaert]

Nice start! Maybe add one of the challenge images and your result ?

[inuex35]

Thanks! Added a Results on PPC-Dataset section with a Tokyo trajectory figure (truth / float / fix + error-colored view) — let me know what you think.

[dellaert]

Some more comments:

• i don’t have to be a co-author - you wrote this ;-)
• Is there is a reason you want to remain anonymous? A real name with affiliation (or, independent researcher) would be preferable
• Spelling mistake in formula (comma after lambda)
• If N variables are assumed known, make them square nodes. If not, how do we find them?
• could you add the formulas for the actual factors (both unary and binary) ?
• I can’t make heads of tails of the figure caption - does not seem to correspond to figure

[inuex35]

Thanks, Frank! Addressed all points:

1. Co-authorship — Removed; single-author now.

2. Byline — Real name: Kosuke Inoue, independent researcher. Personal project.

3. Formula typo — Fixed; now λ·(N_ref − N_target) (no stray comma).

4. Square vs. round N nodes — They're unknown, so kept round. Added how we find them: estimated as floats in the graph, fixed by LAMBDA outside it, then held via fix-and-hold (tight prior or constant substitution), reset on cycle slip.

5. Factor formulas — Added explicit residuals:

   • pseudorange (unary): e_P = Δ∇ρ(x) − Δ∇P̃
   • carrier phase: e_Φ = Δ∇ρ(x) + λ·(N_ref − N_target) − Δ∇Φ̃

   Note it's actually ternary (pose + both ambiguities), matching the code.

6. Caption vs. figure — Revised the factor-graph figure and its caption so they match (one shared reference + two targets; each carrier-phase factor connects the pose and both ambiguities, each pseudorange factor is unary). Could you clarify which figure you meant, though — the factor graph or the Tokyo results figure? Want to make sure I fixed the right one.

Let me know if you'd like more changes!