Our group focuses on research, development and implementation of advanced digital signal processing algorithms, mainly in the fields of adaptive control and audio processing. We build on our experience with statistics, namely with the Bayesian approach to system identification and modeling, as well as the relevant fields of linear algebra.
Our target platforms are Field Programmable Gate Arrays (FPGAs) and Digital Signal Processors (DSPs). We use Matlab/Simulink to specify, model and verify algorithms that we later convert and synthesize to HW. As such specialized solutions are likely to be used in embedded systems, we also research features that result in extremely fast execution, use small amount of memory, small chip area or low power consumption. We achieve this both through designing new or modifying existing DSP algorithms, and using architectural properties such as dynamic reconfiguration of FPGAs...
Listen2Future WP7 Podcast recordings
Presentation at Hora Informaticae seminar June 13, 2023
Our colleague Jiri Kadlec will have a talk about HW Accelerated AI Inference and Fast, Recursive QR System Identification. See HORA INFORMATICAE scientific seminar on June 13, 2023, 14 – 15:30 hours at the Institute of Computer Science of the CAS.
Our colleague Raissa Likhonina succesfully presented a poster entitled Fast Bayesian Algorithms for FPGA Platforms at the DATE 2023 PhD forum on 17 April 2023. The PhD Forum is a poster session hosted by the European Design Automation Association (EDAA), the ACM Special Interest Group on Design Automation (SIGDA), and the IEEE Council on Electronic Design Automation (CEDA). It is a great opportunity for young people to present their work to a broad audience in the system design and design automation community from both industry and academia.
Presentation at the Embedded World 2023
Two Edge AI demonstrators were presented by our team during the Embedded World, 14 – 16 March 2023, at the Trenz Electronic booth (Hale 2, Stand 2-209). First demonstrator was an Edge AI person-posture-detection with Full HD video I/O. Second demonstrator was an Edge AI face-detection implemented with reduced-size data processing unit on Zynq Ultrascale device with minimal programmable logic area on low cost Trenz Electronic board. See some photos from the booth.
Listen2Future Kick-off meeting
The EU reseach project Listen2Future has started 1 February 2023. The project is financed by investments from industry, subsidies from the individual participating countries and the KDT JU program. The Kick-off meeting 1 and 2 February at Infineon in Villach was attended by all high-level project partners and also by Yves Gigase, the KDT JU Executive director. See Infineon Press Release.
EFECS 2022, the European Forum for Electronic Components and Systems was held 24 - 25 November 2022 in Amsterdam. At the StorAIge booth, our department had opportunity to show the latest demonstrator of edge AI pedestrian detection.