By Dr. Fakhar Khalid, the Chief Scientific Officer at Sensat
Infrastructure projects are incredibly complex, involving many different expert teams, often resulting in siloed work. At the same time, this age-old industry has been slow to the technological transformation we see in other industries. However, the onset of the pandemic highlighted for many infrastructure stakeholders that embracing technology must move from theory to practice. The architecture, engineering, and construction industries (AEC) are still in the early stages of adopting AI technology. Yet, this evolution can mitigate risks, improve communication and productivity, and help meet sustainability targets.
AI applied to infrastructure construction
Any infrastructure project involves various key players, from the stakeholders to the people working on the ground. Technological transformation is positively impacting every person and function within a project.
Examples of this include the potential for autonomous vehicles and machines to reduce risk to human workers, disruption to the project, and increase speed and efficiency for many tasks. A construction site is a viable testbed for this technology as it is a contained, safety regulated and a process controlled environment (less unpredictable than a public road, for example). So there are benefits to both the construction industry and the development of this technology for further applications in the future.
The use of algorithms is becoming integral to the planning and managing of complex infrastructure projects. Traditionally, it was challenging to predict required material quantities with complete accuracy. Therefore the safer option was to overestimate and over-subscribe to ensure shortages did not create delays and other issues. Since humans remove an estimated 100 billion tonnes of raw materials from the earth’s fabric each year, and a third of solid waste created globally is attributed to the construction industry, more accurate estimation is critical for the future of zero waste construction.
Algorithms can improve accuracy by surveying past project data to provide estimations with more speed and efficiency than project leaders within siloed disciplines. Technological firms are partnering with construction stakeholders to apply AI algorithms in this manner. Currently, ‘narrow intelligence’ can do this with increasing effectiveness as more data is fed in. In the future, as this technology progresses, unsupervised machine learning will problem solve from the data it is fed, providing further benefit to project estimations.
In design, Computer Vision is used to understand architectural designs. Site monitoring provides an analysis throughout a project that can be compared against the designs to measure accuracy and provide qualitative reports. Algorithm based automated Clash Detection can highlight deviation from the design quickly, preventing significant errors. These technologies positively impact ground workers, project managers, and stakeholders.
These multiple examples of Artificial Narrow Intelligence, working in different levels of construction sites, will eventually combine and accumulate to provide a holistic view of an entire infrastructure project. This is when AI can provide predictions with far higher levels of accuracy than humans could envisage alone. Augmented Reality will provide visualisations of the data collected that AI can do qualitative and quantitative analysis on, and every individual involved with the project will be able to access and feed into the digital twin of the physical asset. This will break down the silos that affect communication and productivity.
Impact on a human level
Post industrial revolution, when a new wave of technology and introduction of robots significantly impacted the manufacturing industry, we were ill-prepared to support people negatively affected by jobs and skills becoming obsolete. We have learned from the past as we face this new technological revolution. Although algorithms and machines will almost certainly replace human jobs, there is still time to mitigate the human impact. We are a long way from AI being able to work autonomously without human guidance and intervention. The vision is a beautiful augmentation between technology and humans. Workers in construction industries must be gradually re-skilled in the management and coordination of technology and the ability to work alongside it.
Possibly the most positive impact AI will have on the human level in construction is the benefits to risk management and health and safety. Clearly evident in emergency response applications, using robots can significantly decrease the risk to human life. Sending a robot with a live video feed into a damaged building to assess risk and locate survivors after an earthquake reduces the risk to human emergency responders. Those responders can then develop a more efficient plan to rescue the trapped and injured with negligible level of risk. Construction is a dangerous profession. In the future, AI robots and autonomous vehicles could reduce the number of injuries and deaths due to accidents and human error. Monitoring and predictive technology could also assess potential risks with greater accuracy, allowing them to be reduced ahead of time.
At a stakeholder level, AI adoption will provide comprehensive visibility over infrastructure projects that have previously caused roadblocks to cost, planning, productivity, and communication. Decision-making throughout the management chain will be data driven rather than guesswork and estimation, and this will also improve collaboration between all key players. In the same way that the financial industry has benefited from using algorithms and subsequently caused a boom in fintech, construction tech is being heavily invested in, in order to serve the industry needs. This will encourage a wave of technology startups geared toward providing AEC industry solutions.
Accessibility to existing data and generation of real time data combined with accurate predictions will affect broader elements around the AEC industries, such as the accuracy of an insurance assessment, measuring environmental and social impacts of projects, and more.
AI in infrastructure and the future of cities
AI adoption in infrastructure does not end when the project is complete. These technologies have some exciting potential applications for how we interact with our urban environments on a larger scale.
A city is like a complex organism. Increasing the data we monitor in terms of human requirements and the different systems that support this will allow for better diagnostics for improving efficiency and sustainability. Suppose smart technology is integrated into future infrastructure and retrofitted into existing infrastructure. An entire city or town can provide the data for algorithms to apply the same benefits of risk mitigation, efficiency, and predictive capacity on a vast scale. AI automation can also support the function of sustainability initiatives and green technology by providing the data needed to measure and improve environmental and social impact.
Focusing on the benefits of further AI integration to the individual, data collection has been a cause of concern for many. An ethical grey line has been crossed when this data is used to target individuals and groups for political influence. Still, when it is used to facilitate people’s lives and improve their well-being, the advantages often outweigh the concerns. In public spaces and private homes, sensors can provide convenience, such as alerting the occupant that the freezer has broken, where energy efficiency could be improved to lower bills, or removing the need to check fire alarms manually. They could also alert the emergency services if a fire or other hazards are detected. More importantly, the rise of smart meters have the ability to reduce energy consumption by influencing efficient use of energy. If this technology becomes the norm in every home and human environment, governing bodies will gain a far more accurate picture regarding decision-making.
We are on the cusp of a significant change in how infrastructure is planned, built, and managed through technological integration. This transformation will gradually filter outwards into vast areas of human life and potentially create a better socioeconomic balance. Data visibility will give individuals more control over their environments, such as their energy usage and safety. But crowd-sourced data will also assist governments in identifying patterns and improving issues to benefit wider society.
Dr. Fakhar Khalid is the Chief Scientific Officer at Sensat, leading the AI Research. He is a spatial scientist with a strong interest in computational intelligence, Spatial Cognition, 2D and 3D Computer Vision with a love for Meta-heuristic Algorithms. Prior to Sensat, Khalid conducted machine learning implementations at SparkCognition and completed a stint at NASA’s Jet Propulsion Laboratory, Birkbeck College, University of London, and the University of Greenwich where he was also a senior lecturer in Spatial Analytics and Remote Sensing. He is a frequent keynote speaker and published academic in the fields of spatial fuzzy computing based machine learning and 3D computer vision. Khalid also leads the SensatUrban project; the largest photogrammetry based Open 3D urban data, available for computer vision research.