Tag
Agility Robotics 发布 Digit 5 人形机器人，以安全性和实用性为优先，可在仓库中搬运 23 公斤料箱、9 分钟快充续航 90 分钟，目标是规模化替代高强度重复性体力劳动。
Flexport 介绍了其位于芝加哥郊外的最新自动化履约中心，展示仓库机器人在物流仓储中的落地应用。
亚马逊在全球仓储中心部署超过100万台机器人，并计划继续扩大机器人劳动力，但劳工团体和研究者指出，与人类协作的自动化机器人可能加剧工作强度与安全隐患。文章探讨了在人机协作环境下，仓库工人的未来处境。
The author argues that legs in humanoid robots for warehouse tasks may be overrated, pointing to Tutor's Cassie robot with wheels as a more efficient alternative.
Eluna is a graph-guided, multi-agent framework for automating warehouse standard operating procedures, using asymmetric episodic distillation to fine-tune a smaller model that matches or exceeds larger baselines and achieves 94% expert agreement on ticket processing.
A feature article explores how recent advances in AI, including reinforcement learning and foundation models, are driving the development of general-purpose autonomous robots for workplaces and homes, featuring insights from robotics leaders at Boston Dynamics and Agility Robotics.
This paper presents an offline reinforcement learning framework for optimizing SLAM throughput control in warehouse fulfillment environments, balancing throughput maximization with downstream stability. The approach is algorithm-agnostic and demonstrates that the CQL policy improves system health by 22.97% and reduces throttling duration by 3.18%.
This paper presents a comparative study of Bayesian Contextual Bandits, XGBoost, and Linear Regression for real-time sorter diversion optimization in e-commerce warehouses, showing BCB achieves 2.03% reward uplift with superior online learning and inference latency.
Amazon has deployed over a million warehouse robots, with a nearly one-to-one ratio to employees; Universal Robots launched AI Trainer that can train robots without programming; humanoid robot costs have dropped 40% in a year, starting at $5,900; Automate 2026 will showcase products available for immediate ordering and a humanoid robot forum.
China Post is using AI robots to sort parcels, achieving 1,200 packages per hour and 6.5 million parcels per day at the Guangzhou center, demonstrating the growing real-world application of warehouse robots.
Figure AI's F.03 humanoid robots, powered by Helix-02 neural network, autonomously sorted 249,560 packages over 200 hours without hardware failure, approaching human-level efficiency.
Figure AI announces that their humanoid robots have accumulated 200 hours of operation handling packages, a milestone in real-world deployment.
Researchers from MIT and Symbotic developed a deep reinforcement learning system that optimizes robot traffic in autonomous warehouses, achieving a 25% throughput gain over existing methods by dynamically prioritizing robots to avoid congestion.