Professor Lukumon Oyedele is the founding Director of the Big Data and Artificial Intelligence Lab (BDAL) at the University of the West of England (UWE) Bristol. His research focuses on how emerging digital technologies can enable transformative improvements in productivity and performance in the UK construction industry.

Onsite workers often rely on written or printed materials to help with the onsite assembly of building components. These instructions, created throughout design and manufacturing, can be disjointed and poorly written, with last-minute changes being missed, hindering right-first-time assembly and resulting in an up to 30% loss in productivity (refs: 1 and 2). If they are using cloud-based, intelligent models via Building Information Modelling (BIM) software, workers spend more time accessing and navigating these systems than completing the actual assembly tasks (ref: 3). Current BIM interfaces are too slow and distracting to be practical for safe and efficient onsite operations (ref: 4).

This project will develop an Augmented Reality-based BIM Assembly Sequence Generator (AR-BIM) to assist onsite workers in finishing assembly tasks 50% faster, decreasing assembly costs by 70%. By allowing workers to easily and practically interact with BIM systems, AR-BIM could eliminate the need for paper instructions. As it works directly from the cloud-based model any last-minute changes will be incorporated. Ultimately this will help to ensure a right-first-time assembly; decreasing rework and the overall cost of built assets.

The adoption of BIM has had a transformative impact on the design and manufacturing stages of construction projects, enabling approaches such as Design for Manufacturing and Assembly (DFMA). Professor Oyedele and his team felt that, by integrating a diverse range of digital technologies, BIM could also have a similarly transformative impact in onsite assembly.

With funding from the Transforming Construction Challenge the BDAL team sought to develop an AR-based BIM Assembly Sequence Generator (AR-BIM) to enable faster, collaborative, right-first-time onsite assembly. Taking a multidisciplinary research approach they are integrating several disruptive technologies, such as Conversational Artificial Intelligence (AI), Augmented Reality (AR), Building Information Modelling (BIM), and Big Data, into one practical tool.

The productivity gained by using BIM in the design and manufacturing phases of the construction process can be lost by onsite assembly not also being treated as a manufacturing environment (ref: 5).

The AR-based BIM Assembly Sequence Generator (AR-BIM) will allow onsite workers to literally talk to BIM systems and receive simplified step-by-step audio and visual assembly instructions.
Advanced Natural Language Processing (NLP) and Conversational Artificial Intelligence (AI) will let onsite workers verbally communicate with BIM systems and quickly get the information they need during assembly. Augmented Reality (AR) goggles, such as Microsoft's HoloLens, will complement the verbal instructions and feedback workers get from the BIM system with accompanying visual guidance.

The project engaged with Costain, Bentley, and WSP to gauge their thoughts on AR-BIM and how it could be integrated into their workflows. From this, it was discovered that potential partners were keen to work with iterative products, rather than expecting a perfect solution.

The Transforming Construction Challenge (TCC) has helped the project make several useful connections. Being part of ICURe Lean Launch programme allowed it to testing its proof of concept with potential customers and investors.  It has connected with a potential collaborator in XYZ reality, part of the TCC-funded HoloSite project, which is creating is the world’s first Engineering Grade Augmented Reality Headset. It also used the Value Toolkit developed by 120 organisations, led by the Construction Innovation Hub.

This project aims to accelerate the adoption of improved tools for the onsite assembly of pre-manufactured building components, made possible by BIM and advances in Conversational Artificial Intelligence (AI), Augmented Reality (AR) and Big Data.

Written or printed instructions don't just hinder right-first-time assembly and result in a 30% loss in productivity (ref: 1). They are incredibly frustrating for workers to use.

Advanced Natural Language Processing (NLP) and Conversational Artificial Intelligence (AI) will let onsite workers verbally communicate with BIM systems and quickly get the information they need during assembly.

The productivity gained by using BIM in the design and manufacturing phases of the construction process can be lost by onsite assembly not also being treated as a manufacturing environment (ref: 5).

Assurance
AR-BIM will help to enable error-free assembly through accurate and easy to understand step-by-step verbal and visual guidance. Working directly from the cloud-based BIM systems ensures that, unlike paper instructions, any last-minute changes are incorporated.

Diversity
The AR-BIM has been working with onsite workers from a diverse range of backgrounds to ensure the AI recognises a range of accents, creating an inclusive tool. 

Safety
Unlike current screen-based BIM interfaces, AR-BIM is hands-free, relying on verbal communication. Using AR goggles, such as Microsoft's HoloLens, for visual instruction can also be done in a way that is less distracting or obscuring than traditional screens.