by Richard McPartland
This year's James Forrest lecture looked at the impact Artificial Intelligence (AI) could have on the built environment sector.
It's a timely topic as the industry is being urged to speed up project delivery, reduce out-turn cost and cut its carbon footprint. AI - the ability of machines to learn and think for themselves - has the potential to contribute significantly to these improvements.
The lecture is held each year in honour of the former secretary of the Institution of Civil Engineers (ICE), James Forrest (1859-1895), and takes on a subject concerning the interdependence of science and engineering.
The event is now available to watch in video form online and we round up some of the key takeaways here on theNBS.com.
View the Artificial Intelligence in the built environment lecture
What is artificial intelligence (AI) and why is it important to the built environment?
When we program a computer we tell it exactly what to do. If things go wrong, it's our fault, as programmer and not the machines. AI is about programming the machine to learn, to actually understand large amount of data and make recommendations or even change things itself.
Artificial Intelligence will transform the built environment in the coming years. BIM has provided the building blocks in terms of standards needed to collect data. Computer systems that can analyse large volumes of data to highlight patterns in the performance and use of existing infrastructure assets will enable better decisions to be made on the type of infrastructure that is required in the future and how it can best be delivered.
AI will likely reduce the number of incremental steps that are required to take infrastructure designs through to operation. This will translate to time and money saved in manufacturing construction products and in building and maintaining our infrastructure networks.
What does this mean for design, construction and facilities management?
In terms of design: Behavioural and environmental data collected via a wealth of sensors will enable us to optimise designs and influence behaviour. Datasets across projects, across organisations, will add to our collective knowledge and result in greater efficiency. Algorithms will start to play a greater role in architecture. Huge quality assurance will be needed to ensure that decisions made or influenced by machines are sound. How construction disciplines and clients engage will change - as we move towards exploring virtual worlds and decisions before ground is ever broken. The focus will be on creating spaces online, manipulating digitally, and walking through them with stakeholders.
In terms of construction: Robotics will increasingly take care of the dangerous or fiddly tasks. They can be relied upon to do things right and, in the right conditions, faster, impacting productivity. Deep learning techniques will help us learn how to best to deploy the technologies. The whole process of construction will be re-imagined - from the earliest drawings to the cutting of ribbon and beyond. Continuously improving the data will become the norm and projects will run end-to-end using a wealth of data as we are starting to see already in the Chinese and Indian markets.
In terms of managing the built environment: Deep learning will move us towards intelligent monitoring and management. Smart traffic systems will keep vehicles moving adapting signals and signs to real-world conditions. Road surfaces can be repaired before they fail.
In terms of maintenance: Better data allows us to predict when maintenance or replacement required prior to failure. Moreover, we can expect to take advantage of autonomous tech in inspection and maintenance.
What issues do we need to grapple with to put AI to work?
- Depth v scope? We need to determine where we will get most short term value – a deep model of something or a shallower model across the whole infrastructure.
- How we will get data of sufficient quality to feed the process? Current infrastructure is often old and records poor. In some respects existing data standards make combination hard. Can we afford a massive effort to drive radical improvement - a Google Earth for infrastructure?
- Static or predictive? To what extent do we monitor and act based on what’s happening as opposed to what’s going to happen?
- What about collaboration? While we can work with the world’s biggest and best, their algorithms are proprietary. How do we influence these or even develop our own built on our own sector's needs?
- Map of value versus map of mischief? – How do we protect data once assembled? More data, more participants can add value but also create potential for problems.
What will the engineer of the future look like?
Engineers will increasingly be called upon to think about how we conceptualise problems and put machinery to work. However intelligent a programmed device is (or seems) there'll always be a requirement for human insight and expertise on combining tools and technology. The engineer of the future will be flexible, be ready to go back to school, to keep enquiring and be open to change. Most of an engineer's existing skills will still needed. Change has always been a constant but the new challenge is in working with machines to help us do our jobs better. The more you understand your business, your subject, the more you can contribute to this process. Age and experience are valued, so too are fresh thinking and approaches.
What to read next...
What construction jobs will look like when robots can build things
Breakthroughs in virtual and augmented reality technology are expected to change the worksites of the future. Research scientist George Quezada muses on the construction jobs of tomorrow.
BREEAM app puts eco-data in the palm of your hand
The new UK BREEAM Finder Tool allows you to locate your nearest BREEAM certified building and BREEAM Assessors or Accredited Professional.
Creating the 'smart cities' of the future
We explore the practical realities of the modern smart city and how a range of technological innovations can help improve quality of life.