Research map / Large language models / Agents, code and math
Open-Ended Instruction Realization with LLM-Enabled Multi-Planner Scheduling in Autonomous Vehicles
Jiawei Liu, Xun Gong, Fen Fang, Muli Yang, Bohao Qu, Yunfeng Hu, Hong Chen, Xulei Yang, Qing Guo
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 changed | Where | Measured | Effect |
|---|---|---|---|
| replace multi-query LLM decision (DiLu++) with a single-query LLM-generated planner scheduling script | LLM+MPC instruction realization · POINT benchmark | Instruction Realization: 0.51 -> 0.84 | Improved |
| use explicit MPC planners with LLM scheduling instead of diffusion-based test-time optimization (Diffusion-ES) | instruction-realization methods · POINT benchmark | Instruction Realization: 0.28 -> 0.84 | Improved |
| add traffic context to the LLM input (ablation) | Intent inference with DeepSeekV3 · POINT benchmark | Intent Recognition: 0.78 -> 0.86 | Improved |
| add LLM planner scheduling over a single fixed planner (ablation) | MPC planner suite · POINT benchmark | Instruction Realization: 0.17 -> 0.84 (vs Lane Keeping PL) | Improved |
| add LLM planner scheduling over a single acceleration planner (ablation) | MPC planner suite · POINT benchmark | Collision avoidance: 0.57 -> 0.99 | Improved |
| scale up LLM size for intent recognition (ablation) | Intent inference · POINT benchmark | Intent Recognition | Improved |
| add reasoning mechanisms and larger context windows to the LLM (ablation) | Intent inference · POINT benchmark | Intent Recognition | Improved |
| increase LLM inference latency from 0 to 4 seconds (ablation) | scheduling framework · POINT benchmark | Instruction Realization (REA): 0.84 -> 0.39 | Hurt |
| increase LLM inference latency from 0 to 4 seconds (ablation) | scheduling framework · POINT benchmark | Collision avoidance: 0.99 -> 0.98 | No clear effect |
| add a specialized-AD comparison: ours versus PDM-Closed on expert progress | scheduling framework · POINT benchmark | Expert trajectory progress: 0.92 -> 0.82 | Hurt |
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.
- DiLu++ — DiLu: A Knowledge-Driven Approach to Autonomous Driving with Large Language Models (ICLR 2024) · better on 3 of 4 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
- LaMPilot: An Open Benchmark Dataset for Autonomous Driving with Language Model Programs (CVPR 2024 · similarity 0.88)
- HCRMP: An LLM-Hinted Contextual Reinforcement Learning Framework for Autonomous Driving (NeurIPS 2025 · similarity 0.87)
- LMDrive: Closed-Loop End-to-End Driving with Large Language Models (CVPR 2024 · similarity 0.87)
- Drive-R1: Bridging Reasoning and Planning in VLMs for Autonomous Driving with Reinforcement Learning (AAAI 2026 · similarity 0.87)
- AutoVLA: A Vision-Language-Action Model for End-to-End Autonomous Driving with Adaptive Reasoning and Reinforcement Fine-Tuning (NeurIPS 2025 · similarity 0.87)
- AutoManual: Constructing Instruction Manuals by LLM Agents via Interactive Environmental Learning (NeurIPS 2024 · similarity 0.86)
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.
