AI-based vision for automating cable handling and insertions in manufacturing with robots

基于人工智能的视觉，用于使用机器人在制造过程中实现电缆处理和插入的自动化

• 批准号: 10032961
• 金额: $ 42.98万
• 依托单位: CAMBRIAN ROBOTICS LIMITED
• 依托单位国家: 英国
• 项目类别: Collaborative R&D
• 财政年份: 2022
• 资助国家: 英国
• 起止时间: 2022 至 无数据
• 项目状态: 已结题
• 来源: https://gtr.ukri.org/projects?ref=10032961
• 关键词: AI; based; vision; automating; cable

Robots have increasingly been used in the manufacturing industry since the 1960s to perform predictable and repetitive tasks. Cambrian have recently succeeded in harnessing Artificial Intelligence (AI) technology to transform robotic automation in manufacturing, intelligently recognising and picking components and performing assembly tasks. This process involves 'training' the system to recognise a component and then using this 'intelligence' to recognise an individual item in a collection of many items, and determine how best to grasp it. The robot can then proceed to perform a predefined task or tasks with the component. Whilst this technology is revolutionary, there are some tasks that the system can not handle - the primary one being the handling and insertion of flexible cables. It is estimated that 50% of all manufacturing assembly tasks involve cable handling, in industries including automotive, electronics, consumer goods, medical and aerospace, equating to 1.5m people worldwide performing cable assembly tasks.This project is to develop the AI robotics technology so that it can recognise, pick and route cables on a production line for use cases across these industries. This development is complex and involves researching, developing and training multiple AI models and integrating this learning into a usable software product. There is no current commercial state of the art product on the market - Cambrian are unique in having performed proof of concept research into fully predicting the actual cable geometry, in 3D, and then being able to find the places where it deviates from the planned path. The output of the project will be a mature AI solution that can be rapidly marketed as part of our main product portfolio to existing and new customers. The majority of assembly tasks have moved out of the UK over a number of decades, in the main due to cheaper labour costs in Southeast Asia. This technology will provide the potential for SMEs and startups in the UK to get back into the manufacturing supply chain, creating more UK jobs, generating revenue for the UK economy and improving local social and economic conditions in some of the most deprived areas of the country. This is a rare opportunity for the UK to lead the way in manufacturing technology, transforming the automation of assembly lines and creating new job opportunities, and revenue, for the UK economy.

自20世纪60年代以来，机器人越来越多地用于制造业，以执行可预测和重复的任务。Cambrian最近成功地利用人工智能（AI）技术改造了制造业中的机器人自动化，智能识别和挑选部件并执行装配任务。这一过程包括“训练”系统识别一个部件，然后利用这种“智能”在多个物品的集合中识别单个物品，并确定如何最好地抓住它。然后机器人可以继续执行预定义的任务或与部件一起执行任务。虽然这项技术是革命性的，但也有一些任务是该系统无法处理的-主要任务是处理和插入柔性电缆。据估计，在汽车、电子、消费品、医疗和航空航天等行业，所有制造装配任务中有50%涉及电缆搬运，相当于全球有150万人从事电缆装配任务。该项目旨在开发人工智能机器人技术，以便在这些行业的用例中识别、拾取和布线生产线上的电缆。这种开发是复杂的，涉及研究，开发和训练多个AI模型，并将这种学习集成到可用的软件产品中。目前市场上还没有最先进的商业产品- Cambrian是独一无二的，他们进行了概念验证研究，以3D形式完全预测实际电缆几何形状，然后能够找到偏离计划路径的地方。该项目的成果将是一个成熟的人工智能解决方案，可以作为我们主要产品组合的一部分迅速推向现有和新客户。几十年来，大多数装配任务已经从英国转移出去，主要是由于东南亚的劳动力成本更低。这项技术将为英国的中小企业和初创企业提供重返制造业供应链的潜力，创造更多的英国就业机会，为英国经济创造收入，并改善该国一些最贫困地区的社会和经济状况。这对英国来说是一个难得的机会，可以引领制造技术，改造装配线的自动化，为英国经济创造新的就业机会和收入。