Physical AI 前沿雷达 · 2026-09-28

语言:中文 · EN · JA

每日自动更新的具身智能前沿雷达:八条主线、三语精炼、证据可查。

今日重点

1. Beyond Binary Success: Sample-Efficient and Statistically Rigorous Robot Policy Comparison

Generalist robot manipulation policies are becoming increasingly capable, but are limited in evaluation to a small number of hardware rollouts.

2. Using Non-Expert Data to Robustify Imitation Learning via Offline Reinforcement Learning

Imitation learning has proven effective for training robots to perform complex tasks from expert human demonstrations.

3. TACTIC: Tactile and Vision Conditioned Contact-Centric Control for Whole-Arm Manipulation

Whole-arm manipulation involves direct contact with the environment while the robot completes a task by distributing contact across multiple links as contacts form, slide, and break.

4. ダイフクが物流倉庫の完全自動化へヒューマノイドを開発、2030年に実証実験

ダイフクは、「国際物流総合展2026」において、自動倉庫やAMRを組み合わせた物流倉庫向けの自動化ソリューションを披露。併せて、物流倉庫の完全自動化の鍵となるヒューマノイド開発の方向性についても言及した。

前沿基线(人工策展)

1. ISO 25785-1 (under development): safety for dynamically stable industrial mobile robots

2. EgoScale: scaling dexterous manipulation with diverse egocentric human data

3. Z-1: efficient RL post-training for flow-based VLA models

4. PL-CBF: finite-horizon safety at runtime via parallel fallback rollouts

5. Post-train NVIDIA Cosmos 3 Edge for on-device robot control

6. RD-VLA: implicit test-time compute scaling via latent iterative reasoning

主线分布 (近 7 天入选)

主线 #
基座模型(VLA / WAM) 2
数据引擎(采集 / 清洗 / scaling) 4
训练与自我改进(RL / 后训练) 6
仿真与评测(sim2real / 闭环) 12
边缘与实时性(端侧推理 / 控制频率) 6
安全、权限与合规 1
系统与编排(长程 / 多策略) 5
本体与供应链(硬件 / 成本) 8

近期节奏

合规与标准观察


本页逐条分析由 radar-analyst 起草(claude-opus-5.5),经确定性检查与 agent 复核后自动合并,未经人工复核;标题与数字未经改写。

本仓库为技术观察,不构成投资建议、产品承诺或安全认证结论。论文结论以作者自报为准。

方法论 · 每周汇总 · 历史归档