EVI-027
Evidence
Table 5：G2DP 较 DP collision +11.0、progress +10.5、TTC +6.7；progress/along-route 超 PDM-Closed。
| id | |
|---|---|
| updated | |
| type | evidence |
| claims | CLM-027 |
| source | PPR-2606.26017 |
| observation | Table 5 (tab:deepscenario_metrics_placeholder_nr): DeepScenario closed-loop NR, safety & progress metrics |
| method | drone-recorded urban dataset converted to nuPlan format; 411 feasible scenarios after filtering invalid initializations (initial collisions/off-road); zero-shot nuPlan-trained model; NR mode |
| scope | zero-shot, NR; rule annotations limited, cross-scenario ego property variation; overall score/comfort not reported |
Result
vs Diffusion Planner：Collision avoidance 81.33 vs 70.34（+10.99）、Progress 74.88 vs 64.42（+10.46）、TTC 66.72 vs 60.05（+6.67）、Along expert route 54.36 vs 47.52。vs PDM-Closed：Progress 74.88 vs 68.90、Along route 54.36 vs 46.00 反超；Drivable area 68.48 低于 PLUTO* 70.31；Driving direction 89.54 低于 PDM-Closed 92.21。
Caveats
411 场景规模小且经初始化过滤（幸存者偏差）；NR 模式无反应式交互；“约 +11.0/+10.5” 为论文正文的约数表述，表内精确差为 +10.99/+10.46。
关联（12）
- PPR-2501.15564 Diffusion-Based Planning for Autonomous Driving with Flexible Guidance
- PPR-2504.17371 Highly Accurate and Diverse Traffic Data: The DeepScenario Open 3D Dataset
- PPR-2606.26017 G2DP: Diffusion Planning with Spatio-Temporal Grid Guidance
- FRM-2606.26017 G2DP 五层重建：核心对象是 CPT-013 时空代价体（占据概率栅格 × 路线进度图融合）在推理期的可微引导；公理层为能量重加权目标分布、Top-K 车体聚合与末段注入等六个作者前提；论证主干是 Table 4 在同一 DiT 骨干下隔离『稠密网格 vs 稀疏实例』的引导模态消融（稀疏臂按 PPR-2501.15564 附录 C 实现、以 oracle 邻车供给）。
- CLM-027 零样本迁移真实无人机记录的密集城市交通（DeepScenario，NR）：安全与路线推进增益保持，碰撞避免与推进维度大幅超过 Diffusion Planner、推进维度反超规则基线 PDM-Closed；drivable area 与 driving direction 非最优。
- EVI-024 co-source Table 1：G2DP 系居纯 IL planner 首位，Test14-hard R 77.61（+7.2 over Flow Planner，+8.4 over Diffusion Planner）。
- EVI-025 co-source Table 4：Grid +3.35 score/+3.59 collision/+4.94 TTC；Sparse(Oracle) −1.96 score/−10.13 comfort/+0.64 collision。稀疏臂按 PPR-2501.15564 附录 C 式 10 的实例级有向距离能量实现，与稠密臂在信号形态与信息内容两轴同时不同（评注已补）。
- EVI-026 co-source Tables 2–3：interPlan 零样本 G2DP† 62.55（+9.65），collision +10.15、TTC +10.75；progress/speed limit 略降。
- EVI-028 co-source Fig. 8：steps 8–9 + lambda_t=0.5 峰值 77.61；早窗与 lambda_t≥0.8 均退化。
- EVI-029 co-source Table 6：Top-15 双split峰值 92.26/80.05；Average 86.33 洗信号、Max 76.84 抗伪影差。
- EVI-030 co-source Fig. 7：γ=0.95 峰值；γ=1.0 过度保守刹车致高密度区追尾，γ=0.2 安全弱化。
- EVI-031 co-source Table 1/2 + 正文：G2DP† reactive +3.7、interPlan +0.81、精修依赖 +4.1 vs DP +12.8。