The research profile of this group is centered around different types of communication. The group primarily works with wireless communications applications, theoretical studies of communication and the mechanisms used to transmit information over a communication medium. More specifically, the group activities can be divided into the following topics:
Channel measurements and modeling
Channel measurements and modeling have a long tradition within the group and currently the activities are focused on multiple-input/multiple-output (MIMO) channels and systems with ultra-wide bandwidths (UWB). The group has a unique set of channel sounding devices, with very high performance. Data from measurement campaigns are used to build and adapt models of previously not studied propagation channels. Research results from these activities are used both for international standardization work and for development of new radio transmission methods.
Algorithm development for digital transmitters and receivers
An integrated part of the design of new wireless communication systems is the study of algorithms with low implementation complexity, performing close to the theoretically optimal methods for transmission and synchronization. The group focuses on such algorithms for time/frequency synchronization and detection/decoding.
Transmission theory / Digital communications
Work in this field centers on methods for wireless communication. Bandwidth-efficient methods -- those that carry many bits per hertz of bandwidth -- are of special interest. The group studies new error-correction coding and speech and image coding methods, all with this aim. There are also projects in powerline communication and spread spectrum transmission.
The research within information theory is focused on understanding the underlying factors that determine the error correcting capabilities of convolutional codes. In addition to structural properties of convolutional encoders, development and analyses of efficient decoding algorithms for (block and) convolutional codes are covered. Presently, efforts are concentrated on construction of and decoding methods for codes on graphs. Using graph-based methods a rate R=5/20 convolutional code with overall constraint length 67 was constructed. Its free distance was determined to be as large as 120.