AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

PPAD: Iterative Interactions of Prediction and Planning for End-to-end Autonomous Driving

Zhili Chen, Maosheng Ye, Shuangjie Xu, Tongyi Cao, Qifeng Chen

ECCV 2024 · Label-efficient and robust learning › Recognition and long-tail · approach: Dense features and foundation priors · 28 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
replace one-shot planning with iterative prediction-planning (PPAD) (ablation)VAD-style planner · nuScenesL2 avg (m): 0.72 -> 0.60Improved
add hierarchical key objects attention (ablation)PPAD · nuScenesL2 avg (m): 0.59 -> 0.58 (full vs without)Improved
add key objects attention to PPAD alone (ablation)PPAD · nuScenescollision rate avg (%): 0.23 -> 0.35Hurt
add confidence-aware collision loss (ablation)PPAD with key objects attn · nuScenescollision rate avg (%): 0.35 -> 0.18Improved
add noisy trajectory training (ablation)PPAD with key objects attn · nuScenescollision rate avg (%): 0.35 -> 0.19Improved
add noisy trajectory on top of CA collision loss (ablation)PPAD · nuScenesL2 avg (m): 0.61 -> 0.58Improved
add noisy trajectory on top of CA collision loss (ablation)PPAD · nuScenescollision rate avg (%): 0.18 -> 0.19No clear effect
add map interaction to ego-agent interaction (ablation)PPAD · nuScenesL2 avg (m): 0.61 -> 0.66Hurt
add BEV interaction on top of agent and map interactions (ablation)PPAD · nuScenescollision rate avg (%): 0.28 -> 0.18Improved
interact at every step (6 iterations) instead of 2 or 3 (ablation)PPAD · nuScenesL2 avg (m): 0.65 -> 0.58Improved
more interaction iterations (ablation)PPAD · nuSceneslatency (ms): 306 -> 385Hurt
use PPAD instead of VADVAD · Argoverse2L2 avg (m): 1.81 -> 1.34Improved

Later papers that compared with this method

3 later accepted papers report PPAD as a baseline in their results tables; 3 of them report a better number on most of the results they share with it.

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy