Impairment-aware Signal Detection (ISIDE)

Background

Future communication systems will rely on lower-grade hardware components than current systems, to reduce the size, price, and power consumption. This is important in both advanced multi-antenna transceivers that require very many components and internet-of-things devices that require long battery lives. The drawback of lower-grade components is that these have impairments; a variety of different non-linear and time-varying distortions will affect the communication signals. The traditional communication theory is developed for distortion-free systems, and when applied to real systems, the distortion is typically viewed as additional noise.

Purpose and goal

The project is motivated by the fact that the distortion caused by impairments is not noise, but has a structure that can be utilized to enhance the communication performance. The project focuses on two blocks in a signal receiver: the equalizer that can mitigate distortion and the detector that estimates which information was sent. Mathematical analysis and communication theory will be used to derive new algorithms and insights from refined setups. These insights will be utilized to treat more challenging but realistic setups with many different types of impairments. Expert-knowledge-powered machine learning methods will be utilized to develop benchmarks on how high data rates can be achieved in the presence of impairments, and new algorithms will be developed to provide performance close to these benchmarks.

Expected outcome

The project will demonstrate how the conventional physical-layer communication protocols can be evolved to enable efficient communication under the presence of transceiver hardware distortion. The solution will be based on utilizing the distortion instead of combating it. The new solutions will be particularly important to enable the use of low-cost low-energy devices as well as large antenna arrays with many components.

Planned approach and implementation

The project is led by Professor Emil Björnson and is funded by a Wallenberg Academy Fellow from the Knut and Alice Wallenberg Foundation. A doctoral student has been recruited to work on this research project.

Duration

2021-2025

Contact person

External link

See more information on the Knut and Alice Wallenberg Foundation web page .