Data driven artificial intelligence for construction

by Rob Charlton, CEO Space Group

Since construction started its transition to digital, there has been a growing question; what can we achieve using the data we have collected and can this data transform how we work on a day-to-day basis.

We all understand the importance of data and the need to have quality, consistent data throughout the life cycle of a built asset. But what if this data can be used in progressive technologies and reach beyond its current confines. Given that data is at the centre of Artificial Intelligence (AI), can we then argue that AI has a place in the construction industry.

Several years ago most of us would have said this was not a realistic prospect, however now we see the emergence of the driverless car, therefore surely it’s possible to take this technology and apply it to construction. Whilst opportunity now presents itself, the first thing to do is define AI in the broadest sense, then we can consider how this may affect construction and its activities.

AI is where machines, such as computers exhibit intelligence. Where computers are able to use algorithms to learn and solve problems from the information inputted and received to essentially simulate human thinking. Together with other fundamentals of computer science, such as programming languages, operating systems, distributed systems and networks, this allows AI to work very similarly to the human brain - a deterministic information processing device with distinct algorithms and modules. In other words, free thinking!

An algorithm is a step-by-step sequence of rules that sets out how to make a decision about something, these then tell the computer what it needs to do to make that decision. Initial development stages focussed on mental tasks such as chess or speech recognition, where as more recently, the focus has been on drones, robotics, face detection and so on. This progress has led many in our industry to consider how the data and developing technology in AI can be used to make the process of design, building and operating buildings more efficient.

The most commonly used term when referring to constructing in this area is "machine learning". Machine learning is a type of AI that provides computers with the ability to learn without being explicitly programmed. Machine learning focuses on the development of computer programs that can change when exposed to new data, similar to data mining. Both search through data to look for patterns. However, instead of extracting data for human comprehension as with data mining, machine learning uses that data to detect patterns in the data and adjust actions accordingly.

Already we see that developers are considering how they can use product and project data, there are areas within our industry which are making progress and those who are working on broader applications, for example there are many projects considering robotics and how this may assist construction. Scaled Robotics received funding to develop mini robots which can construct buildings and there are ideas around the use of drones carrying materials to site and even constructing a wall.

All of this is made possible by the computer taking information from their environment, making thousands of calculations to find the right trends and learn the route needed to optimise delivery.

Other areas ripe for development are Smart buildings and Smart Cities. For example, learning techniques can be used to predict building energy consumption to help with efficiency in automating heating and cooling systems, the usage can be measured with sensors across the entire system to assess performance and energy consumption. As well as reacting to the combined data, the model can predict the amount of energy which will be used or how the system should be calibrated.

We could take this a step further by using machine learning to predict energy usage, and direct power supplies to where they are needed most, preventing any over-supply.

We can also consider the performance of building materials and structure. Sensors on steel frames or concrete will provide data on how a structure is performing. By assessing the data captured there will be the ability to identify the need for maintenance at a specific time rather than a one size fits all maintenance regime.

One of the most existing areas of development is Autonomous site machinery - effectively the driverless car for construction - allowing the driver to be removed from a vehicle in dangerous situations such as working at height or deep excavations. The current technology uses a lead vehicle which links to GPS data. Then from the data collected via sensors, the safest and most efficient route is calculated. This may sound pie in the sky, however Komatsu is already trialing such a system.

These examples only scratch the surface of the opportunity to make our industry more efficient and as we find new ways to collect data the opportunities to use it intelligently are only restricted by our imagination.