Bayesian Active Learning for EMI Near-Field Emission Characterization of High-Speed Electronics

01 January 2024 → 31 December 2027
Research Foundation - Flanders (FWO)
Research disciplines
  • Natural sciences
    • Machine learning and decision making
  • Engineering and technology
    • Wireless communications
    • Electromagnetism and antenna technology
    • Electronic circuit and system reliability
    • Scientific computing not elsewhere classified
Electromagnetic Compatibility Bayesian Active Learning Near-Field Scanning
Project description

The “Internet-of-Things”, “Industry 5.0”, “Smart Cities”, and “Autonomous Vehicles” will bring huge benefits to society. As these technologies become more widely adopted, we will be surrounded by electronic devices that are wirelessly connected. Our lives will become increasingly dependent on the correct functioning of these very complex electronic devices. But, as these devices shrink, take on more functionalities, and are squeezed closer together, the likelihood that they interfere with each other will increase. This is because every electronic device emits electromagnetic interferences (EMI), while at the same time being potentially vulnerable to EMI coming from other devices. We simply must know more about how these devices are affected by EMI so that we can make the devices resistant to their effects. A technique known as “EMI near-field scanning” has shown tremendous potential for characterizing how EMI affects electronic devices. Unfortunately, the technique suffers from significant limitations - mainly it is just too slow. This research project will overcome these major challenges by devising advanced new methods based on multiple antennas and smart Bayesian AI algorithms that can characterize the effects of EMI much faster than we can today. These methods can make a huge leap forward in making our highly connected world more reliable, more effective, and more responsive to our needs.