Prof. Luís Geraldo P. Meloni 


Unicamp - Universidade Estadual de Campinas
Faculdade de Engenharia Elétrica e de Computação
Av. Albert Einstein, 400 - C.P. 6101
13083-970 Campinas, SP
Brasil
e-mail: meloni@decom.fee.unicamp.br
Tel. +55 (19) 3788-3818
Fax. +55 (19) 3289-1395

Courses | Research | Publications | Laboratory


Background

Prof. Luís G.P. Meloni received the Electrical Engineer and M.Sc. degrees both at School of Electrical and Computing Engineering, Unicamp, Brazil, in 1980 and 1982 respectively. He received the Ph.D. degree at Université de Nancy I, France in 1985, under Prof. Michel Lamotte. He is currently Assistant Professor with the School of Electrical and Computing Engineering, Unicamp.

His areas of interest in research include:

q       DSP algorithm optimization. Optimization procedures for efficient algorithm implementation in digital signal processor for next-generation networks: VoIP codecs optimization, echo-cancellation algorithms, media gateway control protocols, real-time kernels implementation including TCP and UDP stacks, quality of service in packet networks and network management software tools.

q       Education software tools for Internet. Research and development of software tools for teaching and learning over the Web. The integration of mathematical software offers a high level of interactivity for the students by means of computations and simulations in the matter under study. The integration of animated graphics in 2-D and 3-D also enhances the course quality and helps the understanding of complex subjects. Course content development which uses the above environment combined with multimedia resources offer a new paradigm in Education, which must be studied and researched.

q       Speech recognition algorithms. Research and development of high performance algorithms of speech recognition aiming real time operation, for applications in several fields such as telecommunications, entertainment, automotive, etc. Research of connected word recognition and continuous speech recognition with applications in the above fields.

q       Audio, image and video coding. Research of a new perceptual model for audio and speech coding. Study and research of efficient implementation of these algorithms in digital signal processor. Efficient coding of multimedia signals is very attractive in the next-generation-networks whereas image and video coding are a potential research area of interest.

His experience in the telecommunications area was also enlarged while working for several companies:

q       As R&D manager for Sid Telecom in the development of various projects such as narrow-band modems and software improvements in the Tropico RS switch;  

q       As R&D manager for a telecommunication research center in Brasília - Cepesc - while  responsible of a team of 22 researchers; where he conducted the vocoder project;

q       As R&D manager for Zetax Tecnologia (now Lucent Technologies) in digital signal processing area in the development of many projects such as MFC R2 transceivers, DTMF detectors, announce machines, telephony pulse detectors, and telecommunications managements software.

He also worked at Brasilia University (UnB) as assistant professor for three years.

 

Research Projects

Multichannel Signaling Algorithms for Telephony. Research and development of multifrequency signaling algorithms for the Texas Instruments TMS320C62x digital signal processor, suitable for multichannel operation. The capacity of 256 channels per DSP is shared by the applications: MF R1 and R2 transceivers, DTMF senders and receivers, announce machines and three-party conferences. The developed algorithms are compliant with ITU-T and LSSGR signaling standards. This project was the object of an R&D agreement with Lucent Technologies and was concluded in December 2000.

Codecs for Voice Over IP. Research and development of highly optimized algorithms for voice over IP codecs. We are developing state of art codecs for the ITU-T G.729A and ETSI GSM AMR standards and for two different DSPs: TI TMS320C62x and Motorola MSC810x (StarCore). The project focuses on the number of channel optimization, memory requirements, and algorithm adaptation to the DSP architecture. This project under development is object of an R&D agreement with Ericsson Telecomunicações.

 

Selected Publications

1. L.F. Melo and L.G.P. Meloni; “Ensino de Engenharia de Tráfego através da Internet”, Revista Digital Colabora, CVA-RICESU, ISSN 1519-8529, October 2002, to apear in the next volume.

2. L.M.J. Barbosa and L.G.P. Meloni; “A Sequential Search Algorithm with Signal-Selected Pulse Amplitudes”, IEEE International Telecommunications Symposium – ITS2002, Natal, RN, September 8-12, 2002.

3. H.C. Bertan and L.G.P. Meloni; “LSP Transcoding between G729A and GSM AMR”, IEEE International Telecommunications Symposium – ITS2002, Natal, RN, September 8-12, 2002.

4. P.C. Feijão, J. J. Bandeira, G.F.G. Yared and L.G.P. Meloni; “Equalizador Digital de Áudio Empregando a MLT”, Seminário de Engenharia de Áudio, UFMG, July 9-12, 2002.

5. R.V. Andreão e L.G.P. Meloni; “Implementação em Tempo Real de um Sistema de Reconhecimento de Dígitos Conectados”, 19° Simpósio Brasileiro de Telecomunicações, SBrT, Fortaleza, CE, September 3-6, 2001.

6. L.G.P. Meloni; “Learning Discrete Hidden Markov Models”, Computer Applications in Engineering Education, John Wiley & Sons, Vol. 8, issue 3, pp. 1-9, October, 2000.

7. I.H.S.P. Fantini e L.G.P. Meloni, “Otimização dos Codificadores VSELP e EFR por Refinamento da Modelagem Autoregressiva”, 18° Simpósio Brasileiro de Telecomunicações, SBrT, Gramado, RS, September 3-6, 2000.

8. I.H.S.P. Fantini e L.G.P. Meloni, “Enhanced VSELP Coding by a Refined Autoregressive Modeling”, 17º Simpósio Brasileiro de Telecomunicações, SBrT, Vila Velha, ES, September 7-10, 1999.

9. L.G.P. Meloni e L.O. Garcia, “Learning Discrete Hidden Markov Models with Emphasis on Speech Recognition”, International Conference on Engineering and Computer Education, Rio de Janeiro, RJ, August 11-14, 1999.

10. L.G.P. Meloni, “Transformadas de Hartley de Radical Dois para Processadores Digitais de Sinais”, 9° Simpósio Brasileiro de Telecomunicações, SBrT, São Paulo, SP, September 2-5, 1991.

11. L.G.P. Meloni, “Algoritmos Rápidos de Extração de Formantes a Partir do Modelo Auto-Regressivo”, 7° Simpósio Brasileiro de Telecomunicações, SBrT, Florianópolis, SC, September 3-6, 1989.

12. L.G.P. Meloni, M. Lamotte and M.J. Vignerion, “Modélisation du Conduit Vocal et Compression Spectrale par Modification du Modèle”, ITBM – Innovation et Technologie en Biologie et Medecine, vol. 7, n° 2, pp. 152-166, 1986.