iTEAM

Home Research Research Areas Signal Processing

Signal Processing

Digital processing of signals applied to: image, sound, communications, sensor data and biological data. Industrial and financial applications.

Research Topics

Applications of biomedicine

We collaborate with the Clinical Area of Medical Image of the Hospital Universitario and Politcnico La Fe (GIBI230) in the processing and analysis of medical images. Providing a long experience in the theory and application of Statistical Signal Processing techniques, automatic detection and classification algorithms, multimodal signal processing, data mining for the extraction of relevant information, detection of novelty, fusion of decisions, higher-order spectral based algorithms for the detection of nonlinearities and mathematical morphology algorithms applied to image processing.

Currently, the main interest is extracting image biomarkers for the quantification of changes associated with the disease. As well as using the available tools to adapt the huge amount of information available in the image environment (DICOM, PACS, RIS) to the effective and real-time control of all quality aspects relevant to service excellence such as: radiation dose, number of procedures per process, temporary adjustment of demand, analysis of large consumers, costs, technological evaluation, etc.

Non-Destructive Techniques (NDT)

Within the research lines, the GTS works on signal processing for nondestructive testing, both for the detection and for the characterization of damage in cementitious materials. Among all the different non-destructive essays, the group activity is mainly focused in the processing of: ultrasonic signals (conventional and air coupled), impact-eco signals, georadar signals, acoustic signals and laser-ultrasonic signals. The developed signal processing algorithms are employed in multiple applications: material characterization, damage location, tomographic reconstruction, acoustic localization ...

The GTS collaborates with the ICITECH (Institute of Science and Technology of the Concrete) for the characterization of different types of global damage in cementitious materials using sonic and ultrasonic signals. For this porpoise wide frequency analytical algorithms combined with signal modality features extraction are employed.

Underwater Acoustics

We employ Passive Acoustic Monitoring techniques to advance in the study of marine mammals and anthropogenic sounds. This is achieved by developing new acoustic instrumentation and signal processing algorithms to detect and characterise underwater acoustic events. Our goal is to improve the understanding of how anthropogenic sounds impacts on the marine biodiversity as well as to study the population abundance, seasonality and behaviour of marine mammals. 

For this purpose, we combine traditional Passive Acoustic Monitoring techniques with recent advances in signal processing algorithms and representation techniques such as big data representations of events automatically obtained by means of deep learning.

High performance arithmetic and DSP operators for FPGA

Design of arithmetic operators and digital signal processing kernels optimized for their implementation in Altera and Xilinx FPGA devices.

FPGA-based DSP for ultrasound waves

Ultrasounds are used in non-destructive measurements in nuclear and aeronautic sectors. This line is focused in the design of digital signal processing algorithms for ultrasound waves and their hardware architectures for FPGA devices: FIR and IIR filters, multirate filters, pipeline-interleaved filters, envelope detectors, logarithmic converters,interpolators for Giga-sample operation, beanforming for array sensors. This R&D line is developed for the company TECNATOM under the contract Desarrolllos de tecnologas electrnicas.

Algorithms and architectures for FPGA-based software radios

Algorithms and architectures for FPGA-based software radios Algorithms and architectures to implement digital communication systems in FPGA devices: re-sampling for transmission, digital up & down conversion, synchronization for QAM systems, synchronization for OFDM systems, digitally implemented analog modulations. Three FPGA-based demonstrators have been developed: 10 Mbps digital IF QPSK MODEM, 54 Mbps base-band OFDM MODEM and fully digital FM receiver.

Algorithms and architectures for advanced Forward Error Correction (FEC)

The objective of this research is the development of algorithms and architectures for hardware implementation of FEC blocks that will be required in future communications systems. We have focused in: Binary Low-Density Parity-Check codes decoders, Non-binary Low-Density Parity-Check codes decoders and Soft decoding of Reed-Solomon codes. We want to improve the operation of the LDPC decoders for high SNRs where the error-floor can appears. We have developed an FPGA-based hardware LDPC emulator to accelerate simulations for very low bit rates.

Signal Processing for Audio Applications

There are multiple audio applications that benefit from signal processing algorithms. Some examples we have expertise on are:

- Spatial directionality of arrays of microphones / loudspeakers. 
- Spatial audiometry using synthesized spatial sounds by software. 
- Evaluation of the perceived annoyance of a sound and the relation to its objective characteristics in order to improve its subjective perception.

Efficient Implementation of Algorithms in Multicore and Manycore Devices

When considering multiple microphones, rate frequencies of 44100 samples/second/input, and a high computational burden due to complex algorithms, efficient software implementation shows to be crucial. Research on how multicore and manycore (GPUs) devices can be used for task parallelization is an open and very promising issue.

Distributed and Collaborative Sound Signal Processing

Different sound applications interact with the environment using one or more transducers but centralizing their associated signal processing. Advances made in the field of distributed computing, together with the availability of the necessary hardware and software, have allowed for the development of powerful sound signal processing systems whose interaction with the environment involves multitude of transducers sets. Applications: Different sound environments in a single space, Human-machine interfaces based on sounds, Seamless mobile sound spaces.

Adaptive Filtering

Adaptive filters are basic in multitude of fields such control, aerospace, communications, automotive, etc. Linear and non-linear filters can be used including a wide variety of adaptive algorithms as LMS, RLS, AP, ... and combinations of them. Applications: Room compensation, active noise control, Noise identification.

Active Noise Control

Low- frequency noise levels can be reduced if a counter-phase wave is added at the listener location. When noise has to be cancelled at multiple listening points, multichannel systems are needed to generate the proper signals. Consequently, advanced signal processing algorithms have to be carefully designed to assure listeners perceive a lower and, possibly, less annoying noise. Applications: Car engine noise, airplane, car and train interior noises.


Contact:

Instituto de Telecomunicaciones
y Aplicaciones Multimedia (iTEAM)
Edificio 8G. Planta 4, acceso D
Universitat Poltcnica de Valncia
Camino de Vera, s/n
46022 Valencia
SPAIN

Email info@iteam.upv.es

Phone +34 963879580

Fax +34 963879583

Follow us

Intranet: Username: Password:
SITE MAP      iTEAM @ 2012    Terms of use      Design o3