Workflow
智平方AlphaBot(爱宝)系列机器人
icon
Search documents
3年1000台!智平方拿下具身智能机器人订单
Zhong Zheng Wang· 2025-09-11 13:25
智平方所采用的人形轮式机器人,无需对工厂场景进行大规模基础设施改造。机器人可在狭窄通道中灵 活穿梭,并精准操作为人设计的工装设备,降低了部署成本与集成难度。 目前,智平方已先后在汽车制造、生物科技、公共服务等领域落地机器人应用,其年产能超千台的自有 工厂已于今年9月正式投产。智平方相关负责人表示,公司今年订单已突破500台。 据悉,双方将以PCB(印制板电路)操作为首个示范场景,共同推动具身智能在半导体显示和智能终端 制造环节的落地应用。智平方AlphaBot(爱宝)系列机器人由端到端VLA大模型驱动,既能凭借大模型 的自适应能力精准完成不同型号、不同位置的PCB操作,又能在工厂的全流程环节中基于同一个本体、 同一款具身大模型在不同场景之间进行其他任务的快速切换,让生产线更高效、更智能。 除PCB操作环节外,机器人还将在OLED真空贴合、耗材管理与尾料回收等多场景发挥作用。 9月11日,具身智能机器人公司智平方宣布与惠科股份全资子公司深圳慧智物联达成合作,在未来三年 内,在惠科全球生产基地累计部署超过1000台具身智能机器人,覆盖从仓储物流、上下物料、零部件装 配到质检测试等全流程。 ...
从“秀技能”到“真干活”:2025机器人商业化破冰进行时
Group 1 - The core sentiment at the 2025 World Robot Conference is the tangible application of robots in various scenarios, contrasting with previous years' focus on skill demonstrations [1][2] - Over 200 domestic and international robot companies participated, showcasing more than 1,500 robot products, marking a record for domestic robot exhibitions [2] - Companies are actively exploring the commercial application of robots in real-world scenarios, with significant orders reported, such as over 2,000 humanoid robots ordered by Songyan Power, primarily for the education sector [3][4] Group 2 - The current capabilities of robots are limited to specific tasks, with advancements needed in their "brain" capabilities to handle more complex scenarios [5][6] - The industry faces challenges such as insufficient "brain" development, limited application scenarios, and manufacturing precision issues, which hinder humanoid robot progress [6][8] - The development of a unified end-to-end model or general model is seen as essential for humanoid robots to advance to higher levels, with significant breakthroughs expected in the next 2 to 3 years [8][9] Group 3 - The industry is focusing on creating a closed loop of "data-model-scene validation" to accelerate the commercialization of robots and expand their application [8][9] - Companies are utilizing innovative approaches, such as combining physical simulation with synthetic data to enhance robot training and performance in real-world environments [9]