Experiment Note
Representative-run plots were added for inter-vehicle gaps, spacing errors, velocities, and accelerations. A trust heatmap visualized the time-varying edge scores and showed periods where attacked V2V edges were downweighted before control.
Aggregate diagnostic plots independently reproduced the primary patterns for RMS spacing error, collision count, and minimum gap. They were described as consistency checks, not as additional independent experiments.
Technical Essay
Aggregate tables answer whether a method wins; trajectories help explain how. Gap and acceleration plots can reveal oscillation, saturation, delayed recovery, or an intervention that a mean RMS value conceals. The trust heatmap connected the learned output to the communication graph and made the edge-level formulation visible.
The representative run was not selected as proof of typical performance. Its purpose was mechanistic interpretation. Statistical conclusions continued to rely on held-out multi-seed aggregates.
PAPER DEVELOPMENT, REVIEW, AND SUBMISSION