lunduniversity.lu.se

Electrical and Information Technology

Faculty of Engineering, LTH

Denna sida på svenska This page in English

Massive MIMO – from prototype to processor

2019-06-19

Illustration of the LuMaMi (Lund University Massive MIMO) testbed serving 4 users.

Massive MIMO* as antenna technology is by now an implemented part of 5G, the fifth generation for mobile communications. The technology allows to use available spectrum much more efficient as any previous technologies. Steffen Malkowsky’s thesis work, presented 17 may in Lund, focused on prototyping and proof-of-concept for what back then was the new technology massive MIMO. We have spoken to Dr. Malkowsky.

“We took the concept for the massive MIMO technology and developed the first real-time prototype worldwide in cooperation with National Instruments and Bristol University. Due to the increased number of antennas, processing complexity scales significantly as compared to previous generations. Moreover, the data have to be shuffled around in the system between the many antennas and the centralised processing blocks”, Steffen Malkowsky explains.

The challenges of complexity and data shuffling were tackled by cleverly distributing the processing within the system. With the designed LuMaMi testbed, they could show that many of the theoretically promised features of massive MIMO did carry over to practice. 

“This is important, as many theoretical results are based on idealistic and simplified models to keep them computable and thus do not capture all the environmental influences in reality. Additionally, performing the first publicly reported test in mobile environments, we showed that even though more challenging, it also works when users do move, for example walking or sitting in a car”, he continues.

“In the end of my project, we used the experience gained during the testbed design in order to develop a processor optimised for this technology to integrate it into a bigger system.

This is important, as a prototype system is specifically designed to provide reconfigurability without having strict requirements on efficiency. However, for actual deployment, metrics such as power and area efficiency are crucial in order to reduce cost and power consumption”, Steffen Malkowsky says.

What made you want to pursue a PhD?
“I did my master’s thesis at the Department for Electrical and Information Technology at LTH, Lund University. I liked the environment, the colleagues and the challenges. Doing research provides freedom to incorporate your own ideas and does not force you to stay completely within one small or well-defined area. When I started I did not really know much about wireless communications, but knowing that Sweden and specifically Lund had a strong presence in the field motivated me even more to get myself into this area.”

What is the most fascinating or interesting with your thesis work?
“Most fascinating with the thesis work has been the collaboration within several different groups and industry we had while designing, building and implementing the testbed and performing the measurement campaigns. It provided an environment allowing much more insight into other fields and areas as well as getting inspiration. Another highlight has been to research on a technology that is already standardized now that my PhD came to an end. I think there is not many researchers that can see things they are working on so fast being realized in practice.”

What aspects of your research do you think will be applied in practice eventually? 
“The code base developed during testbed design was the basis for the NI MIMO Framework**, a software platform marketed and distributed by National Instruments and used by many researchers over the world.”

What are your plans?
“For now my plan is to stay at the university continuing to do research, and especially working more on our processor design. However, I also want to get involved in some new areas such as real-time processing for machine learning and robotics. What will be afterwards, I cannot tell for now”, Dr Steffen Malkowsky concludes.

 


* MIMO = Multiple (antennas) In, Multiple (antennas) Out. Massive MIMO ≈ very many antennas…

**Read more about the NI MIMO Framework here:https://www.ni.com/sv-se/innovations/white-papers/16/mimo-application-framework-white-paper.html

 The Thesis Massive MIMO: Prototyping, Proof-of-Concept & Implementation for download: https://portal.research.lu.se/portal/files/63091253/Thesis.pdf