Site:                Centro Studi E Laboratori Telecomunicazioni (CSELT)
                      10148 Torino
                      Via G. Reiss Romoli, 274

Date Visited: 30 April 1999

WTEC Attendees: N. Moayeri (report author), A. Ephremides, R. Pickholtz, W. Stark, L. Young

Hosts:            Enrico Buracchini, Mobile Services and Radio, Radio Systems, Techniques and Radio Methodologies
                     Giovanni Colombo, Mobile Services and Radio, Head, Mobile Services
                     Gaetano Francesco Cazzatello, Mobile Services and Radio, Propagation, Antennas
                     Claudio Mattiello, Mobile Services and Radio, Propagation, Modeling and Simulations
                     Agostino Moncalvo, Mobile Services and Radio, Head, Satellites and Microwave Radio Relay Links
                     Valerio Palestini, Mobile Services and Radio, Head, Radio Systems
                     Federico Tosco, Head, Mobile Services and Radio


CSELT has 1,223 employees, of whom about 826 work in technical areas and have university degrees. The salaries at CSELT are competitive with those offered by the private industry. CSELT is the research arm of Telecom Italia, which provides 75% of CSELT's budget. CSELT's 1997 revenues and investments were $163 million and $28 million, respectively. If Telecom Italy merges with Deutsche Telekom, there will be only one research center. The mobile activities will be at CSELT, and the German side will handle the network aspects. CSELT collaborates closely with CNET in France. There are also collaborations with British Telecom.

Telecom Italia has 126,381 employees and $26.677 billion of revenues. It spends $589 million on research. (The last three numbers are from 1996.)

G.F. Cazzatello gave a presentation on smart antennas. Linear power amplifiers are important for the proper operation of switched beam antennas. Another type of smart antenna is the angle diversity antenna system. In this type of system, the output from several narrow-beam antennas is compared with the output of a wide-angle antenna, and the best signal is selected. That signal is the output of the GSM receiver connected to one of the narrow-beam antennas. The output is the soft decision output of the Viterbi equalizer and before the Viterbi decoding. Ray Pickholtx, a WTEC study team member, commented that one should be able to do better with maximal ratio combining rather than simply selecting one of the signals.

The next type of smart antenna discussed was the adaptive phase array antenna. This antenna tries to reduce interference by adaptively introducing nulls. In the case of adaptive beamformers in the GSM system, there is a 26-bit midamble in each 148-bit frame. This information is used by the adaptation algorithm.

C. Mattiello gave a presentation entitled, "3G Systems Modeling." He emphasized propagation channel modeling in a micro-cellular structure. Typical height of the base station is 3-5 meters. A database of the geographic area is constructed based on aerial photography. This results in a map with a resolution of one meter. The height information for buildings is also incorporated in these maps. The information is then used to create a color-coded map of field strengths in the geographic area. Mattiello showed a profile of delay spreads as a function of spatial position at 1,800 MHz. This was done with a fixed transmitter and a moving receiver in a corridor (perhaps a street). The delay profile in an indoor car-parking garage looked like a mess with many peaks. He went over ray tracing, which is very expensive, and ray launching, which is more tractable but not very accurate. The latter computation was done with 360 rays, one per degree. He showed a comparison between actual signal strength measurements with the ray launching method and with another method taking diffractions into account.

E. Buracchini gave a presentation on software radios. Software radios are not likely to be implemented or used until 2005 - 2010, even if everything goes right. The main obstacles are the need for high resolution ADC/DACs and the complexity of the processing. There was a discussion on how software radios should be introduced. Should there be a common standard, such as Windows CE or Java OS, or should there be proprietary platforms for each provider? What would be best for the manufacturers and users? Most participants felt that a common platform would be superior for all parties involved.

G. Colombo gave a presentation entitled "Network Architecture for 3G Systems." GPRS should be introduced in the beginning of 2000 for GSM. The problem of macro-diversity in CDMA systems was addressed. In a wideband CDMA system, a mobile might be communicating with a few base stations during soft handover. An important issue is how the higher levels of the network should handle the situation. One possibility is to change the core control point in the network. He talked about the use of ATM Adaption Layers to handle both circuit-switched and IP type of connections on an ATM network.

Published: July 2000; WTEC Hyper-Librarian