Assess and improve localization accuracy (post-radar-association)

[jonnyspicer]

Overview

Assess current geometric localization performance and implement improvements if needed, after radar-only association is working.

Current State

• Basic geometric localization using EllipseParametric and EllipsoidParametric
• SphericalIntersection for simple triangulation
• No optimization-based position refinement

Priority Revision

DOWNGRADED from MVP Critical → Post-MVP Enhancement

Reason: Current geometric methods may be adequate for MVP. The real blocker is radar-only association (#20), not localization accuracy.

Target State

• Performance assessment of current geometric methods with real radar-only data
• Selective improvement only if geometric methods prove insufficient
• Levenberg-Marquardt solver if significant accuracy gains are demonstrated

Dependencies

• BLOCKED by Implement RadarOnlyAssociator for pure radar tracking #20 - Cannot assess localization accuracy without radar-only association working
• Should be completed after Implement RadarOnlyAssociator for pure radar tracking #20 and Implement TDOA/FDOA processing for multi-static radar measurements #21 (TDOA processing)

Tasks

• Wait for radar-only association (Implement RadarOnlyAssociator for pure radar tracking #20) to be completed
• Create test scenarios with pure radar data (no ADS-B)
• Assess current localization accuracy in radar-only mode
• If improvements needed, implement LM solver or other optimization
• Validate improved accuracy without performance degradation

Success Criteria

• Current localization accuracy quantified with radar-only test scenarios
• Clear decision on whether improvements are needed
• If implemented, new methods improve accuracy measurably
• Performance remains suitable for real-time processing

Note: This is now a post-MVP enhancement since the real MVP blockers are radar association and TDOA processing.