Eva Maria Herrmann interviews Professor Paul Lukowicz, head of the Embedded Intelligence research department at the German Research Center for Artificial Intelligence (DFKI GmbH)
How artificial intelligence can revolutionise the way we build
It feels like artificial intelligence (AI) is in the news every day. Everyone is talking about it but they all have a different idea of what it is about. Hollywood blockbusters let humanoid machines take power – remember the on-board computer HAL from Stanley Kubrick’s 1968 film “2001: A Space Odyssey”? What was conceived at the time as future technology is now a reality for many different people. Instead of HAL, today’s systems are called Alexa, Siri, Cortana, or the more neutral Google Assistant. AI is therefore present in many more areas of our lives than we think. The technology is used in the finance industry as well as in new mobility systems, or as an aid to medical diagnoses. Reason enough to take a look at where the technology currently stands and sound out the potential for construction.
PROFILE: What exactly does the term artificial intelligence (AI) mean?
Prof. Dr. Paul Lukowicz: Artificial intelligence is the ability of a computer to analyse a large quantity of data and solve problems using very complex mathematical processes. They range from chess computers as opponents or learning more simple tasks such as facial recognition, through to driverless cars or complete production lines. Compared to just data processing, AI can predict scenarios and make decisions without emotions based on pure data. The systems are generally not fixed but can expand their range by means of programmed learning processes. AI can therefore do some astonishing things, but is only as good as the set formulae and algorithms allow.
PROFILE: That sounds promising. Why, then, do people have reservations about the new technology?
Prof. Dr. Paul Lukowicz: There are several reasons for this. On the one hand, the word »intelligence« is misleading as AI has nothing to do with the word that has emotive connotations for humans. Smartness would probably be a smarter word here. On the other hand, the original intention in the 1960s was to develop a machine that behaves as if it had (human) intelligence. Ethics plays a role here: the discussion about loss of control and manipulation. The priority here must be to use the technology in a way that is beneficial to humans and does not harm them. For example with Industry 4.0, integrated sensors in the machine provide a data basis from which the AI can use statistical methods to identify when a unit will fail, when parts need to be replaced or maintenance needs to be carried out. This is preventative and not only when the systems fail. The workforce is also a hot topic. But it’s not a choice between man and machine, they only work together. In our research, we endeavour to use technology in a way that doesn’t replace humans, but is a tool which makes work more efficient, precise and reliable.
PROFILE: One concept that utilises the strengths of humans and machines is the one from your project ConWearDi. What’s the idea behind this acronym?
Prof. Dr. Paul Lukowicz: The ConWearDi (Construction Wearables Digitization) project aims to develop innovative, technologybased tools which link digital construction site processes to the value added chain in the construction industry. Simply put, this means developing both a web platform for digital information exchange between those involved in the construction, as well as the technical components for this. Tools and materials are enhanced to make it possible to integrate minimal wearable computer systems. This means, for example, that sensors on drill machines can save the position of the drill holes using GPS and copy them, or that wearables such as smart glasses or smart watches can record and process realtime information – for the benefit of the design quality construction sequence and cost reduction.
PROFILE: This all sounds like large amounts of data and very Big Brother.
Prof. Dr. Paul Lukowicz: .That is our biggest challenge. Collecting data during the construction process is not intended to monitor the workforce, but rather provide help in support of the workforce and increased quality. AI analyses data, translates it into a readable output and can differentiate between incorrect and correct measurements using statistical methods. A smart watch fitted with sensors can cover a broad spectrum – from use as construction equipment through to measuring work time and construction noise, i.e. to maintain health and safety at work. We have already been able to gather experiences with sensors on the building site with the Smartwerk (smart factory) predecessor project. In this instance a helmet was designed which can calculate the measurements of a space using installed laser distance and motion sensors, take thermography images and measure humidity and room temperature.
With the ConWearDi project we are going one step further and feeding the realtime construction site information such as building progress and environmental conditions into the digital planning and construction process, in order to utilise the intelligence of the system for monitoring processes which accompany the construction. The building site documentation is identical to its digital twin and can be accessed by everyone involved in the construction. Discussions and agreements made on site are also documented, as are employee assignments, material and building site logistics. Nowadays planning is largely digital, but building is analogue and synchronising this discrepancy using AI presents great potential for the construction value added chain.
Photo: DFKI GmbH