REVIEW 29 cited by
PLUTO: Pushing the Limit of Imitation Learning-based Planning for Autonomous Driving
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
PLUTO: Pushing the Limit of Imitation Learning-based Planning for Autonomous Driving
read the original abstract
We present PLUTO, a powerful framework that pushes the limit of imitation learning-based planning for autonomous driving. Our improvements stem from three pivotal aspects: a longitudinal-lateral aware model architecture that enables flexible and diverse driving behaviors; An innovative auxiliary loss computation method that is broadly applicable and efficient for batch-wise calculation; A novel training framework that leverages contrastive learning, augmented by a suite of new data augmentations to regulate driving behaviors and facilitate the understanding of underlying interactions. We assessed our framework using the large-scale real-world nuPlan dataset and its associated standardized planning benchmark. Impressively, PLUTO achieves state-of-the-art closed-loop performance, beating other competing learning-based methods and surpassing the current top-performed rule-based planner for the first time. Results and code are available at https://jchengai.github.io/pluto.
Forward citations
Cited by 29 Pith papers
-
MISTY: High-Throughput Motion Planning via Mixer-based Single-step Drifting
MISTY delivers state-of-the-art closed-loop scores on nuPlan Test14-hard (80.32 non-reactive, 82.21 reactive) at 10.1 ms latency via single-step MLP-Mixer inference and a latent drifting loss that encourages proactive...
-
World Models as Adversaries: Multi-Agent Self-Play Fine-Tuning for Robust Motion Planning
Role-conditioned adversarial world models with counterfactual credit and regret-CVaR self-play improve closed-loop long-tail robustness of autoregressive motion planners while preserving nominal behavior.
-
Shift & Drift: A Zero-Shot Benchmark for Generalizable and Robust Autonomous Driving Motion Planning
A dual-track zero-shot benchmark (DeepPlan semantic shift + actuation-noise drift) reveals imitation planners collapse under novel urban density and correlated noise while an RL planner remains more robust.
-
Who Responds When the Driver Is Gone? A Framework for Human Intent Understanding
Intent2Drive uses a ToM-inspired LLM reasoner on a new holistic intent dataset to infer latent passenger state and planner objectives, improving structured intent metrics while keeping competitive nuPlan closed-loop scores.
-
G2DP: Diffusion Planning with Spatio-Temporal Grid Guidance
A diffusion-based autonomous driving planner that injects gradients from a spatio-temporal occupancy-and-route cost volume into late denoising steps improves closed-loop safety and progress across three benchmarks.
-
G2DP: Diffusion Planning with Spatio-Temporal Grid Guidance
G2DP constructs a differentiable spatio-temporal cost volume from occupancy and route maps to guide diffusion denoising for collision-free trajectories, reporting SOTA closed-loop scores on nuPlan.
-
G2DP: Diffusion Planning with Spatio-Temporal Grid Guidance
G2DP adds dense spatio-temporal grid guidance to diffusion-based motion planning, reporting +7.2 reactive score gains on nuPlan and improved collision avoidance in zero-shot transfers.
-
UniTeD: Unified Temporal Diffusion for Joint Perception and Planning in Autonomous Driving
UniTeD unifies perception and planning in autonomous driving via shared temporal diffusion with TTM and ARS modules, reporting SOTA results on benchmarks.
-
Dash2Sim: Closed-Loop Driving Simulation from in-the-wild Dashcam Videos
Dash2Sim recovers metric geo-referenced 4D scenes from in-the-wild monocular dashcam videos to enable the ROADWork4D benchmark, revealing that current closed-loop planners fail on work zone lane changes.
-
Mosaic: An Extensible Framework for Composing Rule-Based and Learned Motion Planners
Arbitration graphs that decouple verification and scoring from trajectory generation let Mosaic fuse PDM-Closed and FlowDrive* into a new SOTA closed-loop planner on nuPlan.
-
Mosaic: An Extensible Framework for Composing Rule-Based and Learned Motion Planners
Mosaic integrates rule-based and learned planners via arbitration graphs to set new state-of-the-art scores on nuPlan and interPlan benchmarks while cutting at-fault collisions by 30%.
-
Vec-QMDP: Vectorized POMDP Planning on CPUs for Real-Time Autonomous Driving
By decomposing QMDP belief trees into independent scenario trees and vectorizing tree expansion and collision checks, Vec-QMDP plans in milliseconds on CPUs and matches or beats state-of-the-art nuPlan planners.
-
AlignDrive: Aligned Lateral-Longitudinal Planning for End-to-End Autonomous Driving
A cascaded end-to-end driving model conditions longitudinal planning on the lateral path via anchor-based regression and path-conditioned 1D displacement prediction, achieving SOTA driving score of 89.07 and 73.18% su...
-
AlignDrive: Aligned Lateral-Longitudinal Planning for End-to-End Autonomous Driving
Conditioning speed planning on the predicted path and relabeling synthetic cut-ins yields SOTA Bench2Drive scores (DS 89.07, SR 73.18%).
-
LAP: Fast LAtent Diffusion Planner for Autonomous Driving
LAP plans trajectories in a VAE-learned latent space with one- or two-step latent diffusion, beating prior learning-based planners on nuPlan hard scenarios with up to ~10x lower inference latency.
-
Real-Time Rulebook-Aware Nonlinear MPC for Autonomous Driving with Priority-Biased Tiered Slacks
A tiered-slack NMPC matches expert safety/regulatory compliance on Waymo closed-loop rollouts once log-imitation metrics are separated from driving quality.
-
nuTruck: Benchmarking Autonomous Driving Planning for Distributed Electric-drive Trucks
A new closed-loop benchmark for planning on distributed-electric-drive trucks, built on nuPlan scenarios with a nonlinear six-wheel truck model and a rollover-risk metric, finds that car-oriented trajectory planners f...
-
Agent-driven Long-tail Simulation for Autonomous Driving
LLM agents with structured actions can drive interactive long-tail road users in nuPlan, and SemanticPlan shows current planners still fail safety and semantic completion there.
-
Learning from Mistakes: Rollout-Retrieval Lifelong Policy Learning for Autonomous Driving
R²LPL converts recoverable policy mistakes identified in closed-loop rollouts into corrective supervised targets for lifelong policy improvement, reaching SOTA on nuPlan benchmarks with few cycles.
-
ASSCG: Just-Right Gating over Chattering for Fast-Slow LLM Planning in Autonomous Driving
ASSCG is an RWKV-based adaptive gate trained with SFT and GRPO-style RL that makes Query/Cache/Drop decisions for slow LLM guidance in fast-slow autonomous driving planners, improving scores and cutting latency on nuP...
-
CADET: Physics-Grounded Causal Auditing and Training-Free Deconfounding of End-to-End Driving Planners
CADET is a training-free framework for auditing, benchmarking, and repairing spurious correlations in pretrained end-to-end autonomous driving planners using physics-grounded causal methods.
-
Diffusion Forcing Planner: History-Annealed Planning with Time-Dependent Guidance for Autonomous Driving
Diffusion Forcing Planner applies heterogeneous joint diffusion with time-dependent noise and classifier-free guidance on history segments to generate stable, controllable motion plans for autonomous driving on nuPlan.
-
Decision-Making with Lightweight Confidence-Aware Language Model for Autonomous Driving
Lightweight confidence-aware LM distilled from multi-agent CoT demonstrations achieves SOTA success rates on nuPlan benchmark for AD decision-making with low inference latency.
-
LVDrive: Latent Visual Representation Enhanced Vision-Language-Action Autonomous Driving Model
LVDrive improves closed-loop driving on Bench2Drive by adding latent future scene prediction to VLA models via unified embedding space processing and two-stage trajectory decoding.
-
SafeAlign-VLA: A Negative-Enhanced Safe Alignment Framework for Risk-Aware Autonomous Driving
SafeAlign-VLA uses counterfactual safety pairing and anchor-based group relative policy optimization to incorporate negative data for safer VLA-based autonomous driving.
-
RAPiD: Reward-Guided Consistency Distillation of Diffusion Planners for Real-Time Autonomous Driving
A deterministic transformer policy distilled from a frozen DiffusionPlanner via score-regularized optimization and a PDM-based critic runs 8x faster while roughly matching non-reactive nuPlan scores.
-
End-to-end Conditional Diffusion for Realistic and Controllable Visual Traffic Scenario Generation
A single diffusion model jointly predicts critical background vehicles' states and low-level controls, with gradient guidance steering speed, drivable-area compliance, and collision avoidance/risk during closed-loop s...
-
Large Language Model Enhanced Differentiable Trajectory Planning for IoT-Enabled Autonomous Driving
An IL planner with agent-centric data reuse, complexity-aware async LLM semantics, and residual differentiable optimization reports top nuPlan Hard20 closed-loop scores and real-time CARLA-ROS execution.
-
LUNA-AD: Lightweight Uncertainty-Aware Language Model with Lifelong Learning for Autonomous Driving
LUNA-AD introduces a tri-system model with multi-agent hypothesis exploration, distilled lightweight inference, and reflection-driven lifelong learning that claims state-of-the-art success rates on nuPlan benchmarks w...
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.