An L1 or L2 Multi-Constellation GNSS Front-End for High Performance Receivers

Transcription

1 An L1 or L2 Multi-Constellation GNSS Front-End for High Performance Receivers Ramón López La Valle 1, Javier G. García 2, Pedro A. Roncagliolo 3, Carlos H. Muravchik 4 1, 2, 3, 4 Laboratorio de Electrónica Industrial, Control e Instrumentación (LEICI), Departamento de Electrotecnia, Facultad de Ingeniería, Universidad Nacional de La Plata (UNLP), La Plata, Argentina 1 lopezlavalle@ing.unlp.edu.ar; 2 jgarcia@ing.unlp.edu.ar; 3 agustinr@ing.unlp.edu.ar; 4 carlosm@ing.unlp.edu.ar Abstract-We present the design and implementation of a multi-constellation GNSS front-end. This front-end is able to operate in two different hardware configurations: using the L1/E1 band of GPS/Galileo and the L1 band of GLONASS, or the L2 band of GPS and GLONASS. Both of these operation modes can be implemented in the same printed circuit board by replacing only a few components. In the proposed design, the RF signals are down-converted to an intermediate frequency where the GPS and GLONASS bands are separated. Thanks to this separation, a considerable reduction of the necessary sampling rate for the digitalization stage is achieved. This simplifies and reduces the power consumption of this stage of the complete GNSS receiver. Measurements carried out to the implemented prototypes for the two different configurations are presented. The obtained results validate the proposed design. Keywords- GNSS; RF; Front-End; Multi-constellation I. INTRODUCTION The Global Navigation Satellite Systems (GNSS) are satellite constellations that transmit signals which can be used to determine the position and speed of a receiver located anywhere in the world. The most known GNSS is the GPS, which belongs to the USA, and it has been fully operational since In the last years other GNSS have been developed, like the Russian system GLONASS which was recently declared fully operational. The Galileo system, belonging to the European Union, is still in a development process and it is planned to be completely in operation in Finally, the Chinese system COMPASS/BeiDou is working in China and it is expected to provide global coverage in GNSS applications have been continuously expanding and they are progressively replacing conventional methods of positioning, because of their greater precision and versatility. Having a receiver which combines several GNSS signals should lead to an increase in performance, improving the precision and the continuity of the services under adverse operational conditions. A GNSS receiver is basically composed by three main blocks: an RF front-end, a digitalization stage, and a signal processing stage [1]. The front-end takes the signals coming from the antenna and it provides filtering to select the desired frequency bands. Moreover, the front-end gives the necessary gain to properly drive the digitalization stage. Typically, GNSS front-ends are based on down-conversion to an intermediate frequency (IF) schemes, using one or more conversions [2]. The down-conversion allows reducing the frequency of the received signals, which are found in the range of 1 GHz to 2 GHz. In the digitalization stage, the analog output signals of the front-end are translated into the digital domain by means of an analog to digital converter (ADC). Finally, the rest of the receiver tasks are addressed by digital signal processing. Regarding the front-end architecture, managing the received signals in direct form, though at present possible, involves an excessive power consumption due to the necessary high sampling rates, and its cost is still high [3-5]. Another issue of the direct conversion receivers is related to the jitter in the sample clock of the ADC. Even low jitter levels can degrade the performance of the receiver [6]. In a previous work, we presented the design and implementation of a GNSS front-end able to operate in the L1 band or L2 band using the same printed circuit board [7]. Since only the L1 configuration was implemented and tested, in the present work we include a more detailed description of the hardware and the L2 front-end measurements. The proposed design uses a simple down-conversion scheme. This front-end takes the L1 or L2 signals of GPS and GLONASS coming from an antenna and provides the necessary filtering and gain to the desired bands in order to correctly drive the digitalization stage. Regarding this stage, the use of a low intermediate frequency allows utilizing an analog to digital converter of relatively low sampling rate, which implies low power consumption and reduced cost. This front-end operating with a digital signal processing platform, composed for an acquisition board and an FPGA, allows to implement a whole GNSS receiver [2]. This architecture based on the Software Defined Radio concept is versatile [8], since a programmable receiver for testing acquisition, tracking, and navigation algorithms for research and development purposes can be obtained. The rest of the work is organized as follows. In Section II a description of the signals of interest and of the proposed frontend architecture is realized. A description of the front-end design is presented in Section III. In Section IV, satisfactory

2 measurements results obtained with the implemented prototypes are shown. Finally, the conclusions are given in Section V. II. PROPOSED SCHEME As it was mentioned in the previous section the front-end should be able to receive the L1 or L2 signals of GPS/Galileo and GLONASS. Therefore, we include here a brief description of the signals of interest to determine the frequency bands that the front-end has to manage. The GPS satellites transmit signals in three carrier frequencies called L1, L2 and L5. The last carrier is only present since the current satellite generation (block II-F) though it is not totally available. The L1 carrier has a frequency of MHz and it is modulated by two digital codes, known as C/A and P (actually encrypted). The L2 carrier, whose frequency is MHz, is modulated by the P code. An additional signal called L2C was also incorporated since the previous satellite generation (block IIR-M). Both carriers are modulated by a navigation message of 50 bps. The codes C/A, P, and L2C are different for each satellite, which permits the utilization of code division multiple access (CDMA). The C/A code has a chip rate of Mbps, while the P code has a rate of Mbps. Since the modulation type is BPSK, the bandwidth for C/A code is around 2 MHz and for P code is 20 MHz, considering the main lobe of the spectrum. The power received on Earth is about 125 dbm [9]. The Galileo satellites transmit in three carrier frequencies denoted E1, E6 and E5. All Galileo transmitting satellites share the same frequency bands, making use of the CDMA technique. The E1 carrier has a frequency of MHz and it is modulated by unencrypted ranging codes and navigation data, accessible to all users. Since the ranging codes of this Open Service signal has a chip rate of Mbps and the modulation type is BOC (1, 1), the resulting bandwidth is about 4 MHz. The E6 carrier, whose frequency is MHz, is a commercial access signal and its ranging codes and navigation data are encrypted. The E5 signal is an open access signal which has a carrier frequency of MHz and its bandwidth is around 51 MHz. The minimum received power on Earth is 125 dbm [10]. On the other hand, the GLONASS satellites utilize frequency division multiple access (FDMA) and transmit two signals: L1 and L2. The L1 signal is in a frequency range extended between MHz and MHz, while the L2 signal is between MHz and MHz. Even though each satellite has a different assigned frequency, two satellites can transmit in the same frequency if they are found in antipodeans positions. These carriers are modulated by a digital code which has a chip rate of 511 kbps, a navigation message of 50 bps and an auxiliary sequence of 100 Hz. The resultant bandwidth of each carrier is approximately 500 khz. The power received on Earth for these signals is around 127 dbm [11]. Depending on the hardware configuration, the present design focuses on receiving different signals. In the L1 mode, the front-end can be used to receive the C/A code of the GPS L1 carrier, the Galileo E1 Open Service and the GLONASS L1 band. On the other hand, in the L2 mode the front-end is able to receive the L2C signals of GPS and the L2 band of GLONASS. The P code is not considered since it is encrypted and can only be used by authorized users. Table 1 shows the frequency range of interest for both operation modes. It can be observed that the frequency range of interest has a maximum bandwidth of approximately 36 MHz. Therefore, to satisfy the sampling theorem, it is necessary that the sampling rate must be larger than 72 Msps. Even though this sampling rate is achievable with the actual ADCs, it would produce a large amount of information to be processed and stored, and also elevated power consumption. Having in mind a practical implementation with today s technology, these issues appear to be quite relevant. TABLE 1 FREQUENCY BANDS Configuration Band Frequency Range [MHz] L1 Mode L2 Mode L1/E1 GPS/Galileo L1 GLONASS L2C GPS L2 GLONASS 1570 to to to to 1249 Noticing that the portions of the bandwidth between 1580 MHz and 1598 MHz, and between 1232 MHz and 1243 MHz do not contain any useful information, we considered the alternative of separating the GPS/Galileo and GLONASS bands before addressing the digitalization. In this case, a significant reduction of the used sampling rates can be achieved, simplifying the digitalization stage in relation to the processing requirements, storage of information and power consumption. The proposed heterodyne scheme consists of an RF stage with a low noise amplifier (LNA) which sets the noise figure of the whole front-end. Then, the RF signal is down-converted to an intermediate frequency, by means of a mix with an external local oscillator (LO). After this conversion, the GPS/Galileo and GLONASS bands are amplified and separated. A front-end block diagram is shown in Fig. 1. Fig. 1 Front-end block diagram

3 A. Frequency Plan In order to reduce the distortion produced in the desired signal, the image frequency must be sufficiently attenuated before doing the mixing by filtering the RF signal. If a too low IF is chosen, the image frequency is near the central frequency of the RF signal, requiring an extremely selective filter to achieve sufficient attenuation (superior to 30 db) [7]. A selective filter at these frequencies can be costly and implies large insertion losses. Another issue to keep in mind when choosing the intermediate frequency is related to the bandwidth of the ADC used in the digitalization stage. The ADCs should have a bandwidth compatible with intermediate frequency signals. A high IF requires a wide bandwidth ADC, which implies a high cost. Therefore, for the selection of the intermediate frequency there exists a trade-off between the attenuation of the image frequency and the ADC cost. Considering the attenuation that can be achieved with a commercial filter and the bandwidth of the ADC, an intermediate frequency of approximately 45 MHz constitutes a reasonable choice. The needed local oscillator frequency can be obtained as f LO f f (1) RF where f RF is the center of the band between 1570 MHz and 1606 MHz for the L1 mode, or the center of the band between 1222 MHz and 1249 MHz for the L2 mode. Therefore, the local oscillator frequencies turn out to be 1545 MHz for L1 and 1195 MHz for L2. With these values of local oscillators, the image frequencies are 1502 MHz and MHz for the L1 and L2 bands respectively [12]. These image frequencies can be easily attenuated more than 30 db using commercial filters. On the other hand, the bandwidth required for the ADC is 61 MHz [7], which is compatible with the present technology. B. Gain Calculation The necessary gain depends directly on the power required to drive the ADC of the digitalization stage, i.e. the output power of the front-end. Most of the ADCs suitable for this type of applications have a dynamic range of 1 Vpp (peak to peak). Then, the necessary output power of the front-end is approximately 4 dbm [7]. The signals received by the antenna are immersed in noise, whose power can also be superior to the expected signal level. To determine the received noise power, the thermal noise equation is used [12] IF Ni kts B (2) where k is the Boltzmann constant, T s is the equivalent noise temperature of the antenna in K, and B is the bandwidth in Hz. The noise temperature depends on the type of antenna and typical values are around 100 K [13]. Considering that the antenna s bandwidth is approximately 40 MHz, the noise power received is in the order of 102 dbm. Then, the expected noise power is superior to the signal and it is the noise who determines the maximum gain to avoid saturation of the ADCs. Therefore, the required gain of the front-end is given by [7] G RF 106dB G [ db] (3) where G ant is the gain of the antenna. If the front-end operates with a passive antenna (G ant 0dB), a total gain requirement of 106 db is needed. In case of using an active antenna, which amplifies the received signal and generally has a gain of about 25 db, the necessary total gain is approximately 81 db. Since the objectives of the design contemplate the possibility that the front-end can be operated in both scenarios, active and passive antenna, a gain of at least 106 db must be provided. ant III. DESIGN DESCRIPTION In this section, the design of the main stages of the proposed front-end and the chosen components for its implementation are described. A. Low Noise Amplifier The objective of the LNA is to set the noise figure of the front-end, and for this reason it is directly located next to the antenna input. GNSS signals are too weak because they are coming from satellites located about km from the Earth. In this way, the noise figure is a fundamental parameter of the front-end, since it defines its capacity to detect weak signals. The LNA used in this front-end is designed to obtain a low noise figure and also to give a high gain. The amplifier is based on the low noise transistor BPF740 [14]. It has an input matching network which reflects a value of impedance in the input of the transistor that minimizes its noise figure. It also has an output network that matches the characteristic impedance of the system to the output impedance of the transistor in order to achieve a higher gain. The objective of this design was a minimum gain of 18 db and a noise figure lower than 2 db. The schematic of this LNA is shown in Fig

4 Fig. 2 LNA schematic The designed LNA can be easily modified to operate at the L1 band or L2 band, replacing some passive components in order to tune the input and output matching networks. Since the LNA is one of the most critical components of the front-end, before carrying out the final design, it was implemented in a separated PCB to be characterized in more detail under the two operation modes. The measurements obtained with this prototype are presented in the following section. B. RF Stage Next to the LNA, a SAW passband filter (BPF) was placed to select the RF band of interest. This filter can produce an attenuation of about 40 db to the image frequency [15]. The chosen filter has a version totally compatible with the L2 band [16]. Due to the required gain, which is elevated, it was decided to use an RF amplifier, besides the LNA. If all the gain is placed in the IF stage, i.e. after the mixer, there can be a risk of oscillations or instabilities during operation. In the same way, a double balanced active mixer was chosen, which gives a gain of around 13 db between RF and IF. Fig. 3 is the schematic of the RF stage. Fig. 3 RF stage schematic The RF amplifier and the mixer are suitable for the frequency range of both operating modes (L1 or L2). Therefore, it is not necessary to replace them, which simplifies the front-end and reduces its cost. C. IF Stage After the mixer, an IF filter was placed. Its main objectives are to reject the undesired intermodulation products generated by the mixer and to reduce the bandwidth to the band of interest, reducing the noise. Therefore, it is a bandpass filter that, according to the frequency plan, has a central frequency of approximately 45 MHz and a bandwidth of 36 MHz. We designed a filter with discrete components, because this frequency range is not standard. Since the IF and the bandwidth of the signals for the two operating modes are similar, the same IF filter can be used. This reduces the number of components to implement the front-end. The received signal power can vary widely depending on the location of the antenna or the type of antenna used (active or passive). Thus, it is fundamental to have the capability of controlling the front-end gain and if necessary making adjustments. Having in mind this requirement, a variable gain stage was added at IF. For this purpose, a variable gain amplifier controlled by voltage was used. In this design the possibility of controlling the gain in a manual form was given, by means of a potentiometer regulation, or in an automatic form. For the last option, a digital to analog converter (DAC) was placed, so that the front-end is able to receive the gain information from the digital processing stage and to convert it to an analogical voltage for the control of the amplifier gain. Like the IF filter, the selected variable gain amplifier is compatible with both operating modes. Two MMIC amplifiers connected in cascade were used in order to obtain the rest of the gain. Since they are the last gain stages, an amplifier with 1 db compression point high enough to manage the necessary output power was selected. These amplifiers have a wide bandwidth so it can be used in both operation modes. The GPS/Galileo and GLONASS bands are separated in IF by a duplexer. This block has one input where the IF signal enters, and two outputs: one for the GPS/Galileo band, with rejection to the GLONASS band; and other for the GLONASS band, with rejection to the GPS/Galileo band. This stage is basically composed of two filters with one common input, where

5 each filter allows the corresponding band to pass and rejects the other. The duplexer design was made with discrete components, since the central frequencies and the bandwidth are specific for this application. The separation between the GPS and GLONASS signals in the L1 band is different from the separation in the L2 band. For this reason, depending on the operation mode, some passive components should be replaced in order to modify the central frequencies of the GPS and GLONASS outputs of the duplexer. The schematic of the IF stage is shown in Fig. 4. Fig. 4 IF stage schematic IV. RESULTS The implementations of the prototypes were made in PCBs of RO4350 material. This material is adequate for RF, since it has low losses in high frequency and its dielectric constant is stable [17]. Also it can be laminated in multilayer printed circuit boards, which is a fundamental characteristic because the design s complexity required a four layer printed circuit board. Fig. 5 Front-end prototype The design criteria used consisted in placing the radio frequency lines in the top layer. In this way these lines became microstrip ones, since a ground plane is found underneath. Another internal plane was used for the power supply connections, separating the different voltages in the board. A reduction of the resistance and inductance of the power supply lines can be obtained with this strategy, avoiding coupling problems between the different circuits of the front-end. The bottom layer was used to make auxiliary connections and to put decoupling capacitors of the different devices in the board. With the objective of minimizing the parasitic components due to the high frequency involved, and to reduce the area of the printed circuit board, it was decided to use surface mount devices (SMD) size Two complete front-ends were implemented: one for the L1 band and another for the L2 band. Fig. 5 is a photo of one of the implemented front-ends. Fig. 6 LNA prototype

6 The LNAs for the L1 and L2 front-ends were characterized using the prototypes built for these purposes. Fig. 6 is a photo of one of the implemented prototypes. The measured S parameters of the LNAs are shown in Fig. 7. The LNAs measurements confirmed that in both cases the gain is approximately 20 db, with a correct matching in their inputs and outputs. The measured noise figure was 1.1 db for the L1 band and 1.2 db for L2 band, these values meet the objectives of design by far and they are considerably below the typical values of the commercial receivers. The power consumption of both LNAs is only 5 ma at 3.3 V. (a) L1 LNA S Parameters. Fig. 7 S Parameters of the LNAs (b) L2 LNA S Parameters Afterwards, some measurements to verify the functioning of the front-ends were carried out. For the L1 prototype, the spectra in the IF outputs of GPS/Galileo and GLONASS were obtained exciting the input of the front-end with tones of MHz and MHz respectively. Particularly, in order to carry out the measurements, tones of 95 dbm were used, setting the front-end with a gain of approximately 85 db. The results obtained show the selectivity of the front-end in each output. It can be seen that the presence of the tones in IF is in the correct frequencies: 30 MHz for GPS/Galileo and 57 MHz for GLONASS, without observing spurious frequencies and interferences. The power difference between the outputs of GPS/Galileo and GLONASS is only 1 db, which shows that the gain in the pass band is approximately constant, as expected. The spectra of the GPS/Galileo and GLONASS outputs for the L1 front-end are plotted in Fig. 8. (a) GPS/Galileo output. Fig. 8 Spectra in the IF outputs of the L1 front-end (b) GLONASS output. For the L2 prototype, the spectra in the IF outputs of GPS and GLONASS were obtained driving the input of the front-end with tones of MHz and MHz respectively. The measurement conditions, i.e. the power of the test tones and the gain of the front-end, were the same used for the L1 front-end. It can be observed the IF tones at 32 MHz for GPS and 50 MHz for GLONASS. As in the L1 front-end, the spurious frequencies are not relevant and the power difference between the IF outputs is less than 1 db. The spectra of the GPS and GLONASS outputs for the L2 front-end are shown in Fig. 9. (a) GPS output. Fig. 9 Spectra in the IF outputs of the L2 front-end (b) GLONASS output

7 The total noise figure of front-ends was measured. For both front-ends the obtained results were very similar. The measured noise figure was about 1.6 db, which is lower than the values that can be typically found in commercial front-ends. Finally, a functional test was realized to both front-ends. In this way, an acquisition of actual GPS and GLONASS signals was made [13]. For this purpose, a passive L1/L2 GNSS antenna was connected to the input of the L1 and the L2 front-ends. The output IF signals were digitalized by means of an acquisition board connected to an FPGA in order to store the generated samples [18]. The sampling rate used was 20Msps. The samples were processed using acquisition routines. With these routines, at the moment of realizing the test, for both front-ends all the visible satellites of GPS and GLONASS could be acquired, confirming that the front-ends worked correctly. To illustrate this, some results are shown in this work. The correlation peaks obtained with the L1 front-end are shown in Fig. 10. Fig. 11 is a correlation peak obtained with the L2 front-end. (a) Peak of L1 C/A code of GPS satellite 2 (b) Peak of L1 k = 7 GLONASS satellite Fig. 10 Correlation peaks obtained with the L1 front-end Fig. 11 Correlation peak for the L2 CM code of GPS satellite 5 V. CONCLUSIONS The design and implementation of a GNSS front-end able to operate with the L1 band of GPS/Galileo and GLONASS, or the L2 band of GPS and GLONASS was presented. These two different hardware configurations modes can be implemented

8 using the same printed circuit board by replacing only a few passive components, which reduces the cost and the number of components needed. The proposed design allows achieving a multi-frequency and multi-constellation front-end. Combining measurements of multiple GNSS and frequencies should increase the performance of the receiver, improving the precision and the continuity of the services under adverse operation conditions. The measurements and tests made to the implemented prototypes validated the design. Therefore, the characteristics of the presented front-end make it a versatile design valuable for the development of high performance GNSS receivers. VI. ACKNOWLEDGMENTS This work was supported by the following institutions: Universidad Nacional de La Plata (UNLP) 11-I-166, Comisión de Investigaciones Científicas de la Provincia de Buenos Aires (CICPBA), and Agencia Nacional de Promoción Científica y Tecnológica (ANPCyT) PICT /0909. REFERENCES [1] J. Bao-Yen Tsui, Fundamentals of Global Positioning System Receivers: a software approach, 2nd ed. New Jersey: John Wiley & Sons, [2] G. Mac Gougan, P. L Normak, and C. Stahlberg, Satellite Navigation Evolution: The Software GNSS Receiver, GPS World, vol. 15, iss. 1, pp , Jan [3] M. L. Psiaki, S. P. Powell, H. Jung, and P. M. Kintner, Design and Practical Implementation of Multifrequency RF Front Ends Using Direct RF Sampling, IEEE Transactions on Microwave Theory and Techniques, vol. 53, iss. 10, pp , Oct [4] Ching-Hsiang Tseng and Sun-Chung Chou, Direct Downconversion of Multiband RF Signals Using Bandpass Sampling, IEEE Transactions on Wireless Communications, vol. 5, iss. 1, pp , Jan [5] L. Tarazona, M. Bavaro, PRECISIO RF Front-End, in Proc. 5th ESA Workshop on Satellite Navigation Technologies and European Workshop on GNSS Signals and Signal Processing (NAVITEC), 8-10, De [6] M. L. Psiaki, D. M. Akos, and J. Thor, A Comparison of Direct RF Sampling and Down-Convert and Sampling GNSS Receiver Architectures, in Proc. ION GPS, pp , 9-12 Sep [7] R. López La Valle, J. G. García, P. A. Roncagliolo, and C. H. Muravchik, A Practical RF Front-End for High Performance GNSS Receivers, in Proc International Conference on Localization and GNSS (ICL-GNSS), pp , Jun [8] P. B. Kenington, RF and Baseband Techniques for Software Defined Radio. Boston, USA: Artech House, [9] IS-GPS-200F, Global Positioning System Directorate Systems Engineering and Integration Interface Specification, [10] European Space Agency, Galileo Open Service OS SIS ICD, [11] Coordination Scientific Information Center, GLONASS Interface Control Document, [12] R. E. Ziemer and W. H. Tranter, Principles of Communications, 5th ed. New Jersey: John Wiley & Sons, [13] E. D. Kaplan, Understanding GPS: Principles and Applications. Boston: Artech House, [14] BFP740 NPN Silicon Germanium RF Transistor, Infineon Technologies, Munich, Germany, [15] MA09582 SAW Filter MHz, Golledge Electronics, Somerset, England. [16] MP01698 SAW Filter MHz, Golledge Electronics, Somerset, England. [17] RO4000 Series High Frequency Circuit Materials, Rogers Corporation, Chandler, Arizona, [18] J. Cogo, J. G. García, P. A. Roncagliolo, and C. H. Muravchik, High Speed Acquisition and Storage Platform for SDR Application Development, in Proc. VII Southern Conference on Programmable Logic, Apr Ramón López La Valle was born in La Plata, Argentina and he graduated as an Electronics Engineer from the National University of La Plata (UNLP), Argentina in He is currently M.Sc. student. He is an Instructor in the UNLP and member of Industrial Electronics, Control and Instrumentation Laboratory (LEICI). His research interests are in RF and microwave design techniques with applications to Global Navigation Satellite Systems receivers for aerospace applications Javier G. Garcia was born in La Plata, Argentina and received the Electronics engineering degree from the National University of La Plata (UNLP) Argentina in He is a Professor in the UNLP and member of Industrial Electronics, Control and Instrumentation Laboratory (LEICI). He is currently involved in research and development of Global Navigation Satellite Systems receivers for aerospace applications. His research interests are in statistical signal processing with applications to GNSS and Digital Communications. Pedro A. Roncagliolo was born in Chivilcoy, Argentina and received the Ph.D. degree in Electronics engineering from the National University of La Plata (UNLP), Argentina. He is currently a Professor in the UNLP and member of Industrial Electronics, Control and Instrumentation Laboratory (LEICI). His research interests are in statistical signal processing with applications to wireless communications and Global Positioning System (GPS). Carlos H. Muravchik he graduated as an Electronics Engineer from the National University of La Plata, Argentina, in 1973, and received

9 the M.Sc. degree in electrical engineering, the M.Sc. degree in statistics, and the Ph.D. degree in electrical engineering, all from Stanford University, Stanford, CA, in 1980, 1983, and 1983, respectively. He is a Professor at the Department of the Electrical Engineering of the National University of La Plata and a member of its Industrial Electronics, Control and Instrumentation Laboratory (LEICI). He is also a member of the Comisión de Investigaciones Científicas de la Pcia. de Buenos Aires. He was a Visiting Professor to Yale University in 1983 and 1994; to the University of Illinois at Chicago in 1996, 1997, 1999, and 2003; and to Washington University in St. Louis in 2006 and His research interests are in the area of statistical and array signal processing with biomedical, communications and control applications, and in nonlinear control systems. Dr. Muravchik has been a member of the Advisory Board of the journal Latin American Applied Research since 1999 and was an Associate Editor of the IEEE TRANSACTIONS ON SIGNAL PROCESSING from 2003 to