HOME        CONTACTS       

Special Session on Signal Processing and Architectures for 5G Wireless Systems

Arnaldo S. R. Oliveira
José Neto Vieira (DETI-UA/IT-Aveiro)

5G systems will enable important and challenging communication scenarios, namely enhanced mobile broadband, massive machine type communications for the Internet of Things, ultra-reliable and low latency communications for real-time and safety critical applications and support for high density of mobile users. These types of communication will require considerable improvements and breakthroughs in terms of data rate, network capacity, connection density, end-to-end latency, spectral efficiency, power consumption, mobility, coverage and cost. Some key technologies that are being evaluated for these purposes include massive MIMO, 3D beamforming, ultra-dense networks, new physical layer waveforms and channel access mechanisms (e.g. full-duplex techniques), usage of millimeter waves spectrum bands and new paradigms in the core and access network architectures (e.g. Cloud Radio Access Network approaches). This Special Section is devoted to the presentation and discussion of advanced signal processing algorithms and novel implementation architectures that will support the development of 5G wireless communications systems. Prospective authors are encouraged to submit original and high quality papers in one or more of the following topics focused on 5G wireless systems:

  • Software defined and cognitive radio approaches
  • Sparse signal processing concepts for efficient 5G
  • Spatial signal processing for 5G
  • Modulation, coding and waveform for 5G
  • Signal processing for new spectrum opportunities
  • Modelling, simulation and profiling techniques
  • Validation and test approaches and tools
  • Processing architectures
    • Specialized and/or reconfigurable
    • Low power
    • (Massively) parallel
  • Prototyping devices and technologies for 5G
  • All(most) digital transceiver architectures
  • Agile and/or wideband enabled transceiver architectures and design
  • Low power and battery-less transceivers
  • Fronthaul interfacing, processing and optimization
  • Demonstrators and testbeds

Reconfigurable Systems and Tools for Signal, Image and Video Processing

Nuno Roma, INESC-ID, Instituto Superior Técnico, Universidade de Lisboa (Nuno.Roma@inesc-id.pt)
Sebastian Lopez, Universidad de Las Palmas de Gran Canaria (seblopez@iuma.ulpgc.es)

The consolidated adoption of complex signal processing systems in a multitude of application domains has pushed the development of advanced processing structures capable of satisfying demanding computational requisites. Image and video processing systems, often involving strict real-time requisites are good examples of this trend. To overcome these requisites, reconfigurable processing structures have been regarded as a highly viable solution, often integrating FPGA accelerators tightly coupled with general purpose processors to form highly efficient heterogeneous computing infrastructures. Thanks to their reconfiguration capability, these platforms can even be dynamically adapted to the instantaneous requisites of the targeted application, making them particularly efficient in the commitment of performance, throughput, energy and power constraints. This special session aims to bring together active researchers who are interested in prevailing issues and prominent challenges related to reconfigurable platforms for signal, image, and video processing systems.

SCRAT: Smart CameRAs design and archiTecture

François Berry, Institut Pascal (francois.berry@uca.fr)
Maxime Pelcat, INSA Rennes, Institut Pascal (maxime.pelcat@insa-rennes.fr)

Smart cameras aggregate video sensing, processing, and communication into a single embedded platform. The smart camera image processing approach has emerged thanks to advances in three key domains: computer vision, image sensors and embedded computing. One of the major advantages of smart cameras is their ability to deal with a wide range of problems, categorized as:

  • Mitigating communication bottlenecks: embedded processing reduces the data communication between sensors and the host system, removing the bottleneck that frequently occurs in real-time image processing.
  • Optimizing performances: smart cameras avoid using a mainframe computer system, producing compact sizes (for unmanned aerial vehicles for example) and low energy consumption systems.
  • Adapting perception to the application: for these applications, a high-speed feedback link between sensing and processing is fundamental.
  • Networking cameras: distributed processing using smart cameras has major advantages over centralized processing systems, since it reduces the amount of transmitted information.

The motivation for the SCRAT DASIP special session is to bring together researchers working on different areas of smart cameras, including embedded vision, and architectures for perception systems, languages, software environments and programming tools for smart cameras, and architectures and protocols for camera networks.

Real-time Hyperspectral Image and Video Processing

Eduardo Juarez Martinez, Universidad Politécnica de Madrid Rubén Salvador, Universidad Politécnica de Madrid Gustavo Marrero Callico, Universidad de Las Palmas de Gran Canaria

Recent advances in sensor technologies for hyperspectral image and video capture devices allow envisioning a horizon where data streaming with higher spatial, spectral and temporal resolution will be available for real-time processing applications in either parallel computing facilities or high performance embedded heterogeneous platforms. In this scenario, challenges to be addressed are the adaptation of existing legacy processing algorithms or the development of brand new techniques whose implementation on state-of-the-art high performance computing platforms meets the stringent performance and energy constraints of biomedical, remote sensing and precision agriculture applications, among others. This special session will accept submissions on all aspects of Real-time Hyperspectral Image and Video Processing. Specially encouraged are those describing actual implementations of real-time hyperspectral applications on manycore, GPU and FPGA accelerator platforms. Submissions describing restructuring of already existing and introduction of new hyperspectral processing techniques conceived for an accelerated execution, as well as holistic system-level performance measurements such as throughput-per-watt, among others, are also encouraged.

Smart Signal and Image Processing Algorithms for Industrial Applications (Chair: Uwe Mönks)

Helene Dörksen, Ostwestfalen-Lippe University of Applied Sciences, Institute Industrial IT (helene.doerksen@hs-owl.de)
Volker Lohweg, Ostwestfalen-Lippe University of Applied Sciences, Institute Industrial IT (volker.lohweg@hs-owl.de)

Signal and image processing in industrial applications necessitates effective algorithms in terms of implementability and real-time processing. Last decades witness a huge growth in medical applications, communication Big Data analytics and in performance optimisation of manufacturing technologies and automatised production systems. A challenging task is to design and adopt signal and image processing algorithms which aim to support technical systems. Complex technical systems create a large amount of data which cannot be analysed manually. Thus, information fusion mechanisms gain increasing importance. Besides the management of large amounts of data, further challenges towards the fusion algorithms arise from epistemic uncertainties and-mostly overseen-conflicts in the input signals. This session addresses aspects of signal and image processing for real-world applications under the constraints of robustness and stability.

Copyright 2010-2017 HEART Steering Committee   
Valid XHTML 1.1 Valid CSS