Measurements are the key to knowledge. This also applies to intelligent vehicles, machines and production systems. But it doesn't stop there. In order to improve the performance level of these systems under all conditions, these measurements should also be used to robustly adapt their control.
Why this research?
Production environments face quite a few changes. They must start manufacturing in a highly adjustable way (lot size 1) at the lowest possible operational cost. Customers expect zero-defect products, absolute flexibility in customization and short delivery times. We’re evolving from mass production towards mass customisation.
Vehicles are becoming increasingly autonomous. This goes beyond self-driving cars, think of fruit picking robots, autonomously flying drones, automated guided vehicles in logistics, autonomous surveillance boats, etc.
Sensors play an important part in both production environments and autonomous vehicles. For instance, they can tell when a machine requires maintenance, monitor the product quality and calculate the efficiency of the production process. They determine the position and speed of a vehicle, distinguish between free and occupied space, differentiate between humans and machines, identify ripe fruits that need picking or pallet boards to be transported. Today, increasingly powerful, small and cheap sensors are being introduced onto the market, which continuously creates new opportunities.
But measuring is only the beginning. The collected data must be interpreted and, based on that, decisions must be taken to improve the production system and vehicle performance. However, in modern production environments these decisions are becoming increasingly complex because one should not only consider a variable production process but also a continuously changing production environment. Similarly, autonomous vehicles need to operate in complex, dynamically evolving environments. They need to adapt their path to avoid static and mobile obstacles, they must slow down near people but keep going when passing static obstacles.
Only by dedicated research in this field, we will achieve smart products (machines & vehicles) and high-performance production environments that are able to keep production activities in Flanders.
Concrete research objectives
We contribute to the development of technology for intelligent monitoring and advanced control systems and for supporting the decision-making process. We zoom in on the following research domains:
Sensing & monitoring
- Glass fibre sensors: robust alternative for measuring, among others, pressure and temperature.
- Self-calibrating accelerometers for inertial navigation
- Predictive condition monitoring, focusing on bearings, gears, motors and electromechanical powertrains.
- Indoor and outdoor localisation systems: AGVs, mobile robot platforms and drones. These typically make use of a variety of technologies. Our research focuses on finding the most efficient solution for improved accuracy, a higher update rate and a better cost efficiency.
- Vision-based monitoring (1D, 2D, 3D; visible and infrared), design of optical systems equipped with freeform lenses and image processing algorithms.
- Self-learning sensors, which automatically adjust to their environment so as to generate more reliable data, configure themselves and perform self-diagnostic tests upon the occurrence of errors.
Control & decision making
- Optimal control thanks to new, robust operating algorithms that are easier to adjust and operate in real time so as to permanently ensure the highest possible quality.
- Context aware control: detection of actual operating context
- Dynamic path planning: for mobile systems to avoid obstacles
- Optimal trajectory generation: planning trajectory taking constraints into account
- Learning control, enabling the system to adjust to variable production processes and anticipate the demand for lot size 1 production.
- Safe learning: guarantees that constraints are respected during the learning process
- Simple learning: increases robustness through use of simple, parametrized models
- Shared learning: from the system's own experience or from that of similar systems in the fleet
- Supporting tools in the decision-making process that help to filter out relevant data (using artificial intelligence).
- Manufacturing companies looking for a strategy to obtain a zero-defect, zero-machine downtime production.
- Manufacturing companies that need technology for developing flexible, automated production lines.
- Machine builders that wish to incorporate condition and process monitoring.
- Software companies that develop tools for designing and testing mechatronic systems.
Concerned core labs
|MOFORM_SBO||Model based force measurements||01/02/2016||31/01/2020|
|COMBILASER_SBO||Combination of laser addtive, laser subtractive and other laser processes for improved functional part characteristics||01/02/2016||31/01/2020|
|ROCSIS_SBO||Robust and optimal control of systems with interacting subsystems||01/03/2016||29/02/2020|
|MONICON_ICON||Monitoring and control of laser melting processes||01/04/2017||31/03/2019|
|MODA_ICON||Model based data analytics||01/01/2018||31/12/2019|
|Smart Connectivity||Proeftuin Smart Connectivity||01/01/2018||31/12/2020|
|Smart Maintenance||Proeftuin Smart Maintenance||01/01/2018||31/03/2022|
|AVCON_ICON||Avoidance of collision and obstacles in narrow lanes||01/02/2018||31/01/2020|
|MULTISENSOR_ICON||Multi sensor design and validation||01/02/2018||31/01/2020|
|MULTISYSLECO_SBO||Multi system learning control||01/04/2018||31/03/2022|
|CONACON_ICON||Context Aware Control||01/04/2019||31/03/2021|
|BEAMSHAPE_ICON||Optical beam shaping for high-productivity/quality laser-aided manufacturing||01/04/2019||31/03/2020|
|PILS_SBO||Product inspection with little supervision||01/07/2019||30/06/2023|
|HySLAM_SBO||A hybrid SLAM approach for autonomous mobile systems||01/08/2019||31/07/2023|
|DGTwinPrediction_SBO||A digital twin for health monitoring and predictive maintenance|
- Atlas Copco Airpower nv
- Bekaert NV (Bekaert Engineering)
- CNH Industrial Belgium nv
- DANA Belgium nv
- dotOcean nv
- Laser Cladding Venture NV
- 3D Systems
- Maintenance Partners Belgium nv
- Materialise nv
- Michel Van De Wiele nv
- Nikon Metrology Europe NV
- Octinion bvba
- Picanol nv
- Punch Powertrain nv
- Siemens Industry Software nv
- Televic Rail nv
- Tenneco Automotive Europe bvba
- The Kobi Company bvba
- Triphase nv
- VCST Industrial Products
- ZF Wind Power Antwerpen nv
Andrei Bartic - Cluster Manager