Integrating Conversational AI and Augmented Reality with BIM for faster and collaborative on-site Construction Assemblage (Conversational-BIM)

传统的建筑方法因生产率低和对经济发展的贡献不足而臭名昭著(国家统计局，2017)。为了提高生产率，建筑业必须转变其建筑方法并采用有效的数字技术(TIP，2017)。BIM的采用改变了建筑设计的方式，并加强了建筑制造技术的实施，如面向制造和组装的设计(DFMA)。然而，现场前线工人采用BIM组装已制造的建筑部件的情况并不存在。这导致在流程的设计和制造阶段使用BIM获得的生产率收益损失(BCI，2016)。现场一线工人与BIM工具对接的时间比他们完成实际组装任务所花的时间更多。目前的BIM接口对于现场操作来说并不可行，因为它们太慢、太危险，而且对现场一线工人来说很分散注意力(建筑新闻，2017)。在此基础上，研究将引入先进的自然语言处理(NLP)和对话式人工智能，使现场一线工作人员能够与BIM系统进行口头交流。装配作业是复杂的，往往由于每个项目的独特性、装配方法的不一致以及项目团队的多样性和更迭而变得复杂。在现场组装操作中，现场一线工人需要快速了解安装建筑组件的程序，以最大限度地减少组装错误，缩短整体项目工期。采用语音沟通的免提装配支持BIM系统，可将一线工作人员花费的时间减少50%。除了提高生产率外，它还将通过预制/预组装建筑部件的逐步组装指南，进一步增强无错误的组装操作。技术的发展有助于轻松采用BIM进行现场组装，这一技术的开发具有巨大的潜力，将彻底改变目前的施工方法。然而，除了目前的BIM界面速度慢和危险外，其他限制包括视觉障碍、分心以及与一线工作人员相关的健康和安全挑战。该项目旨在利用增强现实(AR)为访问BIM系统和安装指南提供视觉支持，而不会妨碍或分散现场工作人员的视线。这将为在线一线工作人员提供准确和及时的信息，以便逐步遵循建筑构件制造安装指南。例如，现场组装工人只需问一句：“嘿，对话--BIM，引导我完成厕所安装”，系统将通过AR辅助口头指令简化组装过程，AR设备将在现场实际部件上覆盖准确的组装步骤说明。值得注意的是，现场资源之间的协调对于提高生产率和保证更快、更安全的组装至关重要(ICE，2018)。因此，该项目将利用先进的人工智能、计算机视觉和增强现实技术来开发端到端的BIM解决方案，以支持现场组装操作。除了提高一线装配工人的生产率外，该项目还寻求消除协调现场活动的繁琐过程，这些活动通常涉及多名工人和机器。因此，AR辅助对话-BIM系统将确保远程专家协调一致地指导一线工作人员并监测项目进度和生产率。

The traditional approach to construction is notorious for poor productivity and inadequate contribution to economic development (ONS, 2017). With the aim of boosting productivity, the construction sector must transform its methods of construction and adopt effective digital technologies (TIP, 2017). The adoption of BIM has transformed the way buildings are designed and enhanced the implementation of building manufacturing technologies such as Design for Manufacturing and Assembly (DFMA). However, the adoption of BIM by onsite frontline workers for assembly of manufactured building components is non-existent. This results in loss of the productivity gain from using BIM for design and manufacturing phases of the process (BCI, 2016). Onsite frontline workers spend more time interfacing with BIM tools than they spend on completing the actual assembly tasks. Current BIM interfaces are not practicable for onsite operations because they are too slow, hazardous and distracting for onsite frontline workers (Construction News, 2017). On this basis, the research will introduce advanced Natural Language Processing (NLP) and Conversational Artificial-Intelligence for enabling onsite frontline workers to verbally communicate with BIM systems. Assembly operations are complex and are often complicated by the uniqueness of each project, the inconsistency of assembly methods, and the diversity and alterations of project team. During onsite assembly operations, onsite frontline workers are required to quickly understand the procedure of installing building components to minimise assembly errors and reduce the overall project duration. The time spent by frontline workers can be reduced by 50% with the introduction of hands-free assembly support BIM system that utilises verbal communication. In addition to boosting productivity, it will further enhance error-free assembly operation through step-by-by assembly guide for pre-manufactured/pre-assembled building components.The development of technologies to aid easy adoption of BIM for onsite assembly has great potential to revolutionise the current approach to construction. However, apart from the slow pace and hazardous nature of current BIM interfaces, other limitations include visual obstruction, distraction and the associated health and safety challenge for frontline workers. This project aims to utilise Augmented Reality (AR) for providing visual support to access BIM systems and installation guides without obstructing or distracting the view of onsite workers. This will provide accurate and just-in-time information for online frontline workers to gradually follow the installation guide of manufactured building components. For example, an onsite assembly worker can merely ask, "hey Conversational-BIM, guide me through toilet installation" and the system will facilitate the assembly procedures through AR-assisted verbal instructions, the AR device will overlay the exact illustration of the assembly steps on the actual components onsite. It is important to note that onsite coordination between resources is vital for boosting productivity and guaranteeing faster and safer assembly (ICE, 2018). This project will therefore exploit advanced AI, computer visions, and AR technologies to develop an end-to-end BIM solution to support onsite assembly operations. In addition to boosting the productivity of frontline assembly workers, this project seeks to eliminate the tedious process of coordinating onsite activities which often involve multiple workers and machinery. Accordingly, the AR-assisted Conversational-BIM system will ensure a coordinated approach for remote experts to guide frontline workers and monitor project progress and productivity.