_DSC2713-EXP60.jpg

Ambient Intelligence

The Ambient Intelligence Research Group realises state-of-the-art digital technologies that makes 'everything' smart, where deemed necessary. We focus on robust and sustainable future ready IT for a ‘smart world’.

This primarily concerns utilising sensor technologies, artificial intelligence and augmented reality. We use these to develop prototype applications related to safety, healthcare, smart industry, sustainable energy systems, animal monitoring, and several other domains. We are not restricted to specific application domains. We are domain agnostic. We focus on bridging novel research knowledge, methods and tools and industry by helping businesses and organisations improve their process efficiency, product quality and processes with the help of data. Our group’s research core lies in data acquisition/IoT (sense), extracting interpretable hidden insights and learning from the measured data (think) and providing feedback to the user (act).

Our areas of expertise

Connected embedded systems

This line of research focuses on making the latest developments in the field of internet-linked embedded systems applicable within existing engineering practice. We develop practical knowledge and prototypes concerned with obtaining data from sensors (measuring), sensor fusion and filtering, data distribution, communication protocols and (embedded) software development.

Applied data and AI

This line of research focuses on how to arrive at ‘situational awareness’ from measured “sensor” data: who is present, what are they doing, what is going on, or even what are the intentions. In addition to data storage and management, this mainly involves (big) data analysis, data mining, statistical and machine learning (ML) systems, ML engineering and decision support.

Augmented interaction

The Augmented Interaction line of research uses various technologies to achieve natural, intuitive interaction. We use cross reality (augmented, mixed and virtual reality), combined with physical representations such as robots, to make the invisible visible and interactive in a natural way. In our view, there is no longer a separation between interaction with computers and interaction with the real world.

AMI lectoren.jpg

Questions?

Please get in touch!

dr. ir. Wouter Teeuw & dr. Jeroen Linssen

Lectors Ambient Intelligence

News from the research group

Our focus

We focus on the following areas:

IoT Platform

We are investigating (in collaboration with our industrial partners) standardised architectures for secure data exchange. These architectures prioritise data monitoring, cybersecurity, and federated systems, addressing the urgent need for robust and reliable data exchange solutions that many businesses are seeking. Within the research group, we have our own IoT platform: IECON. This platform is ideal for experimentation. It is also the basis for a ‘Data Space’ that we use in the energy domain.

IoT-platform1.jpg

Edge Intelligence

Here we focus on embedded systems from a software engineering perspective (e.g. algorithms, embedded software, frameworks and distribution systems), enabling the creation of embedded intelligent solutions instead of the hardware development of embedded platforms. In this sense, our focus is on understanding the different hardware architectures and available solutions that best suit a specific application and so bring intelligence to the software. The question is always what you do ‘in the cloud’ and what you do in the ‘embedded system’.

Edge intelligence1.jpg

Secure IoT

Cybersecurity is a research area that examines security aspects including confidentiality, integrity and availability, in relation to communication and data infrastructures, where cyber-physical systems need to be protected from unauthorised access. In particular for IoT and cloud applications where data is transferred over insecure communication channels such as the internet. Research topics focus on the field of cyber-secure data distribution systems for IoT applications.

IoT security1.png

eXplainable AI

Key to the discussions concerning AI is the trade-off between man and machine: who has the upper hand in such systems used by humans? This can be linked to eXplainable AI (XAI). You need to be able to find out how a learnt model reaches a specific conclusion. Due to their good performance, machine learning and deep learning methods are increasingly being used to classify errors, for example. This means that the explainability of such models is undermined. To combat this, we investigate how reasoning steps in these models can still be explained, and we focus on finding the balance between their performance and their explainability.

explainable ai.png

Data and ML Engineering (DataOps and MLOps)

The value of data and the corresponding learning models lies in the information and patterns that can be mined/extracted/learnt from it. However, the challenge also lies in continuously ensuring that (i) up-to-date, high quality and reliable information is available and (ii) models are updated based on newly available data. But how do we get from data to meaningful information? To help do this we focus on the operational aspects of data and machine learning. That is engineering and validating methods, tools and frameworks to continuously obtain smart data-driven insights in a scalable and robust manner. We want to ensure that data and model management is well organised and that issues such as data and model drift are handled with due care and, where possible, automatically. This is in essence the core DataOps and MLOps: how do we make data and machine learning operationally available and reliable?

dataflow.png

Data visualisation and graph literacy

How do you make information accessible to users in a good way? Only when ‘ordinary people’ use an application or website, read the information correctly, understand what it means for their situation, and become motivated, can it lead to, for example, the desired behavioural change. Here we focus on a good connection between the data visualisation and the data and graph literacy of the target group.

pexels-goumbik-669612.jpg

Changing behaviour

Gamification, or the application of elements from games, is a way to use psychological motivations to motivate people to show certain behaviour. Anyone who has ever tried to learn a language with Duolingo knows how motivating a simple progress bar can be to practise Spanish vocabulary for just a few minutes. We use similar techniques, for example, to motivate patients in healthcare to do their exercises for longer and more often. We use gamification to stimulate better energy-use behaviour among residents. We also apply elements from game design and micro-interactions to augmented reality instructions in the manufacturing industry, to help ensure that employees can perform their tasks with more focus and pleasure.

Augmented reality

Whereas in previous years we were more concerned with Virtual Reality (VR), the scope is increasingly shifting to Augmented Reality (AR): adding information to the real world. In the manufacturing industry, we are not only working on optimising AR instructions in the workplace, but also how we can use AR hardware in combination with artificial intelligence to capture professional knowledge and then feed this back to new employees. There are many projects running in the health domain, for example improving online sessions between physiotherapists and patients through information analysis and visualisation, AR rehabilitation training in the Intensive Care Unit or training anaesthesiologists using VR.

pexels-michelangelo-buonarroti-8728222.jpg

“I see enormous added value in applied research with higher education. Not only because we benefit greatly from it as a manufacturer, but also in a broader perspective. Our industry will soon have professionals from MBO and HBO who will have to help our country move forward. With these technicians of the future, we as Scania, will remain part of the world around us. We are not autonomous, but need our environment and a new generation of technicians, also for sustainability challenges."

Hein van Rietschoten.jpg
Hein van Rietschoten, Maintenance manager at Scania Production Zwolle

Let’s find a solution together

See options

Our researchers

_DSC2713-EXP60.jpg

The research group in the media

  • Jeroen Linssen was a guest together with professor Wouter Teeuw, on the podcast ‘De Dataloog’: the biggest podcast on data and AI in the Netherlands. Listen to the full episode here.
  • ‘We want to make cycling safer and more enjoyable’ – Read an article by U-Today about our Smart Connected Bikes project here. 

Are you about to graduate or looking for an internship? Want to do more in addition to your study and are you interested in research? Become a member of our research group network and join our network of researchers, companies and former students.

Check out your options!

Related research groups