Using ML/AI to fast-track SME construction digitisation adoption to improve UK building productivity

**背景：**英国建筑公司直接受到新冠肺炎的影响，他们的工作量和管理团队受到干扰，材料/劳动力通胀和防寒措施导致成本上升。该项目旨在优化英国中小企业建筑公司，利用机器学习和人工智能应用于现实世界的问题，显著提高生产率和节省成本。**问题：**每年有成千上万的房屋、扩建工程和小型工业项目受客户委托，建筑师/设计师在不同的CAD系统上或手工绘制图纸。然后，建筑公司通常会获得PDF格式的建筑图，用于报价和施工。通常由3-5个建筑商投标同一项目。其结果是，每个项目只有一个成功的候选者，就会产生数百万份投标，浪费数百万GB成本和数小时的投标准备时间。因此，面对这种财务风险，匆忙产生估计，导致不利于随后数字化管理中标者的产出。需要一种软件工具，帮助建筑商直接从建筑计划中轻松确定所有劳动力和材料、所有相关成本，并准备详细的项目计划，反映工作内容，建筑商自己的资源，包括所有相关的过程管理和文档，准备有效地建造大楼。**解决方案：**机器学习和人工智能(ML/AI)，将ML/AI与HBXL现有软件相结合，可以解决这一问题。应用ML/AI项目愿景是加快所有实体(地板、墙、屋顶和窗户)的数字表示的过程，并从平面推断实体类型和几何形状。ML/AI将所有实体数据传递到HBXL现有软件，链接到规范，估计，项目规划和健康安全管理系统。最终的成果可以在HBXL现有的施工管理软件中由整个项目团队共享和使用，并很快将发布基于云的建筑云。**为什么要为这个创新项目提供资金？**这个创新项目满足了政府在产业战略挑战基金建筑部门和人工智能部门交易中概述的双重目标，重点是数字和人工智能方法来设计、施工和管理。这种及时的方法将自动实现较小项目的整个施工过程的数字化，节省每个项目的准备时间，并在实际施工期间产生至少5%-10%的生产率增长(ONS，2018年)，该项目有可能改变数字排除不那么精通技术、时间紧迫的中小企业建设者参与数字方法的方式，提供易于使用的工具，为他们节省大量时间，同时提高交付速度、质量和可持续性。

**Backdrop:**UK building firms have been directly impacted by Covid-19, with their workloads and management teams disrupted along with increasing costs due to materials/labour inflation and Covid-Secure measures.The project aims to optimise UK SME building firms, making significant productivity and cost savings using Machine Learning and AI applied to a real-world problem.**Problem:**Every year hundreds of thousands of houses, extensions and small industrial projects are commissioned by clients, with Architects/Designers preparing drawings on varying CAD systems or by hand.Building firms are then typically provided with building plans as PDFs for quotation and construction, typically with 3-5 builders tendering for the same project. The result is millions of tenders being produced with only one successful candidate per project, wasting millions in £cost and hours of tender preparation time.Consequently, in the face of this financial risk, estimates are produced in haste, resulting in poor outputs to subsequently digitally manage successful tenders.What is needed is a software tool to assist builders easily ascertain direct from the building plan all labour and materials, all associated costs and prepare a detailed project plan, reflecting the content of the work and the builders own resources, including all associated process management and documentation, ready to efficiently construct the building.**Solution:**Machine Learning and AI (ML/AI), combined with HBXL's existing software can solve this problem with funding.The applied ML/AI project vision is to speed up the process of creating digital representations of all entities (floor, walls, roof and fenestration), and infer the entities type and geometry from the plan.The ML/AI will then pass all entity data to HBXL's existing software, linking to specification, estimating, project planning and health & safety management systems.The resulting output can then be shared amongst and utilised by the entire project team in HBXL's existing construction management software and soon to be released cloud-based Construction Cloud.**Why fund this innovative project?**This innovative project addresses Government's twin goals outlined in the Industrial Strategy Challenge Fund Construction Sector and AI Sector Deals, focusing on Digital and Artificial Intelligence approaches to design, construction and management.This timely approach will automatically digitise the entire construction process for smaller projects, saving days of preparation time on each project and produce at least 5-10% in productivity growth during actual construction (ONS, 2018).The project has the potential to transform the way digitally excluded less tech-savvy, time pressured SME builders engage with a digital approach, delivering easy-to-use tools which will save them large amounts of time whilst simultaneously improving speed of delivery, quality and sustainability.