Education

Master's degree projects

Context of the Instrument 2.0 thesis subjects

The availability of massive, low-power processing on portable platforms enables dynamic adaptation of signals and channels in the 5G and post-5G era. Massive MIMO data transport and highly reconfigurable devices are key to exploit this new opportunity. This new approach represents the next quantum leap for our daily life usage of wireless technology.

While the advantage of this additional freedom is very clear for the user, the design and the characterisation of its components still represents a massive challenge, both at the design and at the characterisation level.

We believe that the quantum leap to instrument 2.0 is needed to tackle two major roadblocks for a fast and relevant characterisation of this versatile and complex new technology:

Flexible components require the flexible and fast instrument 2.0.

Currently, instruments are good in obtaining ‘static’ measurements. They are built to perform one test in one operating point for one setting of the device. This is a complete mismatch with the new technology to be characterized:

• As the number of settings and configurations increases exponentially, exhaustive testing becomes too expensive and too time-consuming.  Partial testing is not an option either, it leads to potentially undetected failures and hence a larger production drop-out. This jeopardizes the creation of cheap and reliable products and reduces the sustainability of the production.
• Characterising devices that adapt to their environment calls ‘environment- aware’ instruments. Current instruments fail to provide a test landscape that mimics the real-world operation of the devices, hereby missing key performance indicators for the day-to-day operation of the devices.

We propose to design a test bench top operate the device in a dynamically controlled environment. General applicability for RF and microwave devices, requires that loading impedances, impinging signals and over-the-air interfaces vary dynamically under careful control and adapted to the required performance of the device under test.  

Smart components require the ‘smart’ instrument 2.0.

Currently, instruments are ‘dumb’ devices in the sense that that they always execute the same test in the same way. This is called for by traceability and validation, but currently it hampers the development of ‘smart’ instruments. That is why we propose to think ‘out of the box’:

• ‘Smart’ instruments adapt to the device under test and the user request.
  Using prior knowledge is a key enabler to speedy and informative measurements. Besides expected device behaviour, prior knowledge also contains the specification of the required accuracy of the measurements.  We propose to group all this in a ‘measurement map’. This is a test plan rather than a single test definition.  The map extends the definition of tests to include the expected knowledge of the device. Besides classical figures-of-merit such as bandwidth and power level, this definition should also clearly define what variable operation means, which figures-of-merit are to be measured and what the minimally required accuracy is expressed in quantifiable terms. This definition is then transformed in one initial fast measurement extracting the ‘coarse’ figures-of-merit inline. Accuracy is monitored, but not imposed during this first test.
• ‘Smart instruments’ find their own way in the device capability jungle.
  Comparison of the expected behaviour of the measurement map and the coarse test exposes the behaviour of the device. To
  his initial knowledge is expected to be incomplete and moderately accurate, adaptation of the measurement plan during the measurement will be mandatory to optimize the measurement and reach the demands of the measurement map.  At the end of the continuous on-line controlled experiment, the goal is to have extracted all the measurements in the test plan with a known accuracy at best, or a known variability at least. This requires a new experiment framework, as defining and optimizing the measurement trajectory is a real challenge. We expect techniques such as reinforcement learning to be an enabler here. 

Below is a list of proposed Master’s degree projects at the Department ELEC. 

Seminars

Seminar presentations are organised on regular basis by the department ELEC. Stay up to date by checking the 'News and Events' on the home page.

Educational projects

Designing systems from concepts:

The design of complex systems demands that engineers possess significant set of abstract system-level thinking skills. Engineering students therefore need to be exposed to the art of solving problems systematically and have to learn the limitations and the backsides of ad-hoc methods, to ensure that they should only turn to these methods as last resort alternatives.