WIP: add CLASLIB parity tooling and track C work#4
Draft
rsasaki0109 wants to merge 9 commits into
Draft
Conversation
Owner
Author
|
Added CLAS developer documentation in commit Scope of this docs-only update:
Validation done locally:
This updates the CLAS development/docs surface only. It does not claim CLASLIB-equivalent solver accuracy. |
Add concurrency groups, timeouts, PR triggers for develop branch, and test_ci_workflows.py to validate workflow YAML structure.
- Startup phase variance inflation to reduce code→ambiguity bleed - Separate iono/trop atmos token selection with freshness-first trop - Beta state trop estimation in IF-LC fixed solution - GPS-only DD retry and iono-aware dd_state_transform for NL AR - WL-NL hold improvements with DD constraint persistence - Extensive debug logging for startup epoch analysis
Root cause: navigation.cpp picked orbit/clock/code-bias from non-preferred CLAS networks (e.g. network 1 at tow=230515), causing 0.55m LOS + 1.12m cbias divergence from CLASLIB. Changes: - pickOrbit/pickClock prefer preferred_network_id entries - mergeCodeBiasGroup uses strictBiasScore first-pass to block non-preferred network code bias, with codeFallbackScore fallback - Interpolation before/after prefer preferred network entries - Disable trop estimation in claslib-parity preset (matches CLASLIB tropopt=off; trop state was drifting to unphysical negative values) Result: 30ep RMS 3D 2.77→2.25m, 100ep 5.90→3.90m, 300/300 fixed.
Swept 4/6/8/10/15/20/50σ on 30+100 epoch benchmarks. 8σ gives best combined score: 30ep 2.44m, 100ep 1.98m RMS 3D. 100ep improved from 3.90m to 1.98m (49% reduction).
Swept 2/3/5/10/20/50. Threshold 10 gives best 100ep RMS 3D = 1.34m (was 1.98m at threshold 3). Higher threshold delays integer fixing until position converges, preventing early mis-fixing. All thresholds maintained 100/100 fixed rate.
SIS continuity delta was computed from observation-time interpolated orbit/clock, giving 4x larger corrections than CLASLIB (e.g. G25: -0.327m vs CLASLIB -0.084m). Now uses exact/held SSR epoch values, reducing G25 delta to -0.092m. Remaining 9mm gap is from orbit reference epoch resolution.
rsasaki0109
force-pushed
the
codex/track-c-claslib-parity-wip
branch
2 times, most recently
from
168f69d to
8e08402
Compare
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Current state
Latest parity numbers
Main remaining gap
Validation performed
gnss_ppp