Combined Galileo/GPS architecture for enhanced sensitivity reception

Transcription

1 Int. J. Electron. Commun. (AEÜ) 59 (2005) Combined Galileo/GPS architecture for enhanced sensitivity reception Andreas Schmid a,, André Neubauer a, Henning Ehm b, Robert Weigel b, Norbert Lemke c, Günter Heinrichs c, Jón Winkel c, Jose Angel Ávila-Rodríguez d, Roland Kaniuth d, Thomas Pany d, Bernd Eissfeller d, Günter Rohmer e, Matthias Overbeck e a Development Center NRW, Infineon Technologies AG, Duesseldorfer Landstr. 401, Duisburg, Germany b Institute of Electronics Engineering, University Erlangen-Nuremberg, Cauerstr. 9, Erlangen, Germany c IfEN GmbH, Alte Gruber Straße 6, Poing, Germany d Institute of Geodesy and Navigation, University FAF Munich, Neubiberg, Germany e Fraunhofer Institute of Integrated Circuits, Am Wolfsmantel 33, Erlangen, Germany Received 12 November 2004; received in revised form 19 April 2005 Abstract Several new requirements and challenges are introduced with the transition of traditional navigation applications towards location-based services (LBSs). This paper introduces the HIGAPS receiver concept. Aim of the HIGAPS project is to develop the concept for a combined Galileo/GPS receiver that is specially tailored for LBSs, E-911, and other consumer market applications. After a brief overview of the receiver, the partitioning into analog hardware, digital hardware, and software is outlined. The architecture of the combined Galileo/GPS RF front-end is presented in low-if topology. The digital baseband presents a highly parallel correlation architecture for combined Galileo/GPS reception, allowing fast times to fix for extended dwell times. Parallel digital signal processing combined with aiding data allows single-shot measurements particularly designed for LBSs. Differential correlation further improves the reception sensitivity. The coherently integrated predetection samples are thereby multiplied with the conjugated complex previous predetection samples Elsevier GmbH. All rights reserved. Keywords: Galileo; GPS; Architecture; Enhanced sensitivity 1. Introduction Most GNSS market surveys anticipate an exponentially increasing market for location-based services (LBSs). LBSs, including the US E-911 mandate, however, exhibit distinctly different requirements that cannot be met by traditional navigation receivers. The undoubtedly most important prerequisite for a GNSS receiver in LBS and consumer market applications is very Corresponding author. Tel.: ; fax: address: A.Schmid@infineon.com (A. Schmid). low cost. The HIGPAS concept is therefore based upon newest high integration CMOS technology and establishes a concept with fewest external components. Advanced digital signal processing that is especially designed for low implementation complexity further reduces costs and power consumption. 2. Receiver overview Low energy consumption is particularly important, since most host systems for LBSs are mobile handsets, such as mobile phones or personal digital assistants. The receiver concept therefore allows operation in two distinct /$ - see front matter 2005 Elsevier GmbH. All rights reserved. doi: /j.aeue

2 298 A. Schmid et al. /Int. J. Electron. Commun. (AEÜ) 59 (2005) Table 1. Proposed partitioning of functionalities between analog hardware, digital hardware, and software Analog hardware Digital hardware Software Downconversion to intermediate frequency Downconversion to baseband Assistance data processing Local oscillator Digital clock generation Visible satellite estimation Suppression of the mirror frequency Doppler shift compensation Frequency deviation estimation Analog to digital conversion Despreading Code phase estimation Coherent integration Data bit estimation Noncoherent/differential integration Storage of the correlation values Selection of integration times Data decoding Synchronization lock detection Calculation of the navigation solution modes stand-alone reception mode and LBSs mode. For stand-alone reception, the receiver performs acquisition, tracks the signals and decodes the data overlay. LBSs mode is aided by assistance data as defined in the A-GPS protocols by 3GPP. The receiver only completes a single-shot measurement with an extended acquisition to accurately synchronize to the Galileo and GPS signals. Doppler shift estimation and the data overlay are delivered by the assistance network. The short operation period not only results in very low energy consumption, it also enables quick response times. The third extraordinarily important requirement is high availability of the positioning service. Typical reception sites for LBSs include deep urban, moderate indoor, urban canyon, and inside vehicles without external antenna. This frequently introduces an additional attenuation of 25 db or more on top of free space loss and multipath fading. Assistance data aiding is one means to support the required enhanced reception sensitivity. The architecture of the combined Galileo and GPS receiver also provides an integrated low-noise RF front-end and advanced digital signal processing for enhanced sensitivity acquisition. The receiver concept considers an optimal gain combination of coherent, differential, and noncoherent correlation. A combined antenna is used for GPS and Galileo. In the RF front-end, the signal is down-converted to an intermediate frequency, where the signal is sampled and A/D converted. Filtering to reject out-of-band interferences and to block mirror signals is also applied in the front-end. Supplementary measures for the BOC signals are necessary due to the multiple correlation peaks. Coherent correlation is combined with differential correlation. The proposed partitioning of receiver functionalities between analog hardware, digital hardware, and software is presented in Table RF front-end For the combined Galileo/GPS receiver front-end a low- IF architecture is proposed, shown in Fig. 1. This architecture utilizes the benefits of a competitive, low power and high integration CMOS process. It requires a minimum of off-chip components. The front-end bandwidth is approximately 4 MHz, which comprises the two main lobes of the Galileo BOC(1, 1) signal as well as the main lobe of the GPS C/A code with its two side lobes. This bandwidth is sufficient since it spans most of the signal power of both Galileo and GPS. Compared to a zero-if architecture, the Fig. 1. Proposed low-if front-end architecture for a combined Galileo/GPS receiver with off-chip band-pass filter.

3 A. Schmid et al. /Int. J. Electron. Commun. (AEÜ) 59 (2005) low-if architecture is insensitive to DC-offsets and flicker noise. Both problems can be avoided by AC-coupling between consecutive stages in the front-end. In zero-if GPS receivers DC-offset compensation is a severe problem, since most of the GPS C/A code signal energy is at DC. The main drawbacks of the low-if architecture are its limited image rejection and the stringent I/Q-mismatch requirements. Both issues are relaxed in a combined low-if Galileo/GPS receiver, when the IF is below 8 MHz. For this choice the image lies in the Galileo/GPS-L1 band and mainly consists of thermal noise. In this case, an image reject ratio [ ] AIQ + 1/A IQ + 2 cos(δ ) IMRR = 10 log 10 (1) A IQ + 1/A IQ 2 cos(δ ) of 20 db is enough to reduce the SNR degradation due to the limited image rejection below 0.1 db. In (1) A IQ denotes the amplitude difference and Δ the phase difference between the I- and Q-path. This IMRR can be achieved with moderate limits for the amplitude and phase error in the I/Q-path. For an amplitude error of A IQ = 1 db and a phase error of Δ =4 the IMRR, due to I/Q-mismatch, is approximately 23 db. Fig. 1 shows the detailed front-end architecture of the signal path. The signal received from the antenna is directly fed into the input of the front-end. The signal is amplified by the first low-noise amplifier (LNA) and fed to the off-chip bandpass filter (BPF). From there the signal is fed back into the chip, where it is further amplified by a second LNA. The LNA is followed by a quadrature mixer which downconverts the signal to a low IF of 3.5 MHz. The signal is amplified after the mixer and complex filtered by a polyphase filter H polyphase (f ) = H LP (f f 0 ), (2) which can be thought of as a frequency shifted low-pass filter [1]. The polyphase filter attenuates the image and rejects out of band noise. The shift frequency f 0 equals the IF in this application. The signal is then sampled with 3 bits, which is enough to limit the quantization degradation of the SNR to less then 0.75 db [2]. 4. Coherent predetection Parallelized hardware allows the combined Galileo/GPS receiver concept to support substantially long dwell times for enhanced reception sensitivity together with fast times to fix. This parallel signal processing combined with assistance data aiding enables single-shot measurements. Fig. 2 visualizes the concept for the parallelized digital baseband processing. The bandpass signal at the output of the common Galileo/GPS RF front-end may be represented by the complex-valued, equivalent low-pass signal r ν, the intermediate frequency f IF, and the sample period T s } r bp,ν = R {r ν e j 2πf IF ν T s. (3) After digital downconversion, the result is the equivalent low-pass signal r ν = 2C d ν c ν e j φ ν + n ν. (4) C represents the carrier power, d ν the data signal, c ν the received spreading code, φ ν the time-variant received signal phase, and n ν the complex-valued zero-mean additive white Gaussian noise. Eq. (4) addresses shifts in code phase, signal phase and signal frequency. Amplitude variations and multipath fading would be modeled with a time-variant carrier power C. The complex Gaussian noise n ν has the variance σ 2 n = E{ n 2 }=2N 0 B F, (5) where F is the receiver noise figure and B = 1/T s the passband bandwidth of the anti-aliasing filter. The baseband signal is then despread with the local PRN reference code c r,ν and coherently integrated over the period T c = L T s s μ = μ L 1 ν=(μ 1) L r ν c r,ν+τ mod L. (6) For sufficiently small frequency deviations Δf μ, the phase terms e j φ ν in (4) can be approximated with help of the average frequency deviation Δf μ during one coherent integration interval [(μ 1) T c, μ T c ]. The resulting average phase after coherent integration is denoted by μ. The predetection samples after downconversion, despreading, and coherent integration then result to [3] s μ 2C d μ R rc (τ) sinc(δf μ T c ) e j μ + w μ, (7) μ = μ 1 + 2πΔf μ T c, (8) where R rc (τ) is the circular correlation function between received code c ν and reference code c r,ν. w μ denotes the resulting complex-valued, zero-mean, additive white Gaussian noise with variance σ 2 w = E{ w E{w} 2 }= L2 T c 2N 0 F. (9) New signal processing challenges arise with the transition from GPS to combined GPS/Galileo, e.g. due to the BOC modulation scheme or longer PRN sequences. Therefore, a flexible architecture that enables reuse of communication and signal processing components throughout the different navigation standards is a possible solution for a multi-standard receiver. The received signal can either be correlated with a GPS or a Galileo code. Besides generating an appropriate chipping clock for the respective standard and detecting epoch boundaries, the clock generation and synchronization unit of a combined receiver has to generate synchronization signals (e.g. for switching between navigation standards). The same principle can be applied to the correlation processor for tracking mode, presented in Fig. 3. Even though

4 300 A. Schmid et al. /Int. J. Electron. Commun. (AEÜ) 59 (2005) Fig. 2. Potential digital baseband architecture with differential correlation for single-shot measurements. Galileo BOC signals require additional correlators (veryearly and very-late replica codes besides the early, prompt, and late correlators for GPS C/A), the basic architecture remains the same. It is therefore possible, to form a combined correlation processor architecture, where the earlyprompt late correlator structure is used for GPS and Galileo, while the additional correlators required for Galileo are activated on demand. This allows to switch between GPS and Galileo signals during runtime. 5. Differential correlation Differential correlation relies on multiplying each current predetection sample after coherent integration with the complex conjugated of the, respectively, previous predetection sample after coherent integration [3]. This gives rise to various possible detector test statistics 2 Λ diff,1 = s μ sμ 1, (10) Λ diff,2 = R s μ s μ 1, (11) N/2 2 Λ diff,3 = s 2μ s2μ 1, (12) N/2 Λ diff,4 = R s 2μ s2μ 1. (13)

5 A. Schmid et al. /Int. J. Electron. Commun. (AEÜ) 59 (2005) Fig. 3. Architecture of the combined GPS and Galileo correlator for the tracking process. Differentially coherent integration as shown in (10) (13) is an alternative method to the conventional noncoherent integration used in current enhanced sensitivity receivers Λ std = s μ 2. (14) Differential correlation allows to synchronize the spreading code phase with lower SNR degradation than conventional noncoherent integration. The sensitivity gain arises from the fact that, for appropriate filter characteristics, differential correlation multiplies two statistically independent noise components, while conventional noncoherent integration multiplies two identical noise components. This leads to a lower combined noise variance for differential correlation. Using the definition for s μ from (7), the differential correlation result Ψ = s μ sμ 1 (15) has the mean value E{Ψ}=2C R 2 rc [ d μ d μ 1 sinc(δf μ T c ) sinc(δf μ 1 T c ) e j ( μ μ 1 )], (16) μ = μ 1 + 2πΔf μ T c. (17) Using the central limit theorem, it can be shown that the result after differential correlation converges very fast ( N 4) to a Gaussian-distributed random variable with variances [3] E {R{Ψ E{Ψ}} 2} = (N 1) σ4 w 2 + σ2 w 2 E {I{Ψ E{Ψ}} 2} = (N 1) σ4 w 2 + σ2 w 2 a μ+1 + a μ 1 2, (18) a μ+1 a μ 1 2, (19) a μ = 2C d μ R rc (τ) sinc(δf μ T c ) e j μ, (20) a 0 = a N+1 = 0. (21) For a stable frequency deviation Δf = const., the differential correlation delivers signal components that are all in phase to each other and therefore accumulate with maximal gain E{Ψ} Δf =const. = 2C Rrc 2 sinc2 (Δf T c ) e j 2πΔf Tc d μ d μ 1. (22) When the frequency deviation approaches zero, Δf 0, the differential correlation delivers signal components that are all fully real valued E{Ψ} Δf =0 = 2C R 2 rc d μ d μ 1. (23)

6 302 A. Schmid et al. /Int. J. Electron. Commun. (AEÜ) 59 (2005) In case the residual frequency deviation Δf μ is sufficiently low, just evaluating the real part of the detector test statistic as shown in (11) and (13) further improves the reception sensitivity by rejecting the quadrature noise component. A detailed analysis of differential correlation and a derivation of the respective probability density functions is provided in [3]. The achievable sensitivity is calculated and compared to conventional noncoherent integration for cases with constant frequency deviation, frequency drift, and strong interfering navigation signals. 6. Multipath mitigation By narrowing the chip spacing d [chips] of the early and late correlators in a noncoherent DLL, e.g. 0.1 chip correlator spacing, the maximum multipath error is reduced by a factor of, e.g. 10 and multipath with relative delays of approx. 1 chip or greater is rejected entirely. Multipath beyond 1 chip results from the fact the reflection is outside the support of the auto-correlation function [4,10]. The noise error is proportional d. The basic concept of the E1/E2 tracking correlation technique is to find a tracking point on the autocorrelation function that is not distorted by multipath [5]. Two correlators with a chip spacing of d [chips] are located on the early slope of the autocorrelation function. Another implementation which is different from this basic approach can be found in [6]. This implementation also uses two early correlation functions R E2 (τ) and R E1 (τ) at the points E1 and E2 to set up the following discriminator function D(τ) = η R E2(τ) R E1 (τ). (24) For a given tracking point τ 0, η can be expressed as η = τ 0 d + T c (25) τ 0 + T c with d [chips] being the correlator spacing between E1 and E2 and T c the chip length of code. The multipath performance was simulated with the IfEN Monte Carlo simulator. The multipath generator within the simulator is a general multi-ray model, suitable for either deterministic input (e.g. based on the results of ray-tracing analysis) or it can generate multipath components corresponding to a statistical description. For the distinct dominant reflectors the corresponding rays were generated with a constant delay and variable phase, varying with a Doppler between 0 and Hz relative to the direct line-of-sight. For the diffuse component, a multitude of rays was generated. In order to maintain the constant overall power in the diffuse component, each ray was assigned an amplitude according to a i = 1 10 (CMR/20), (26) N where N is the number of rays in each diffuse component and CMR the carrier to multipath ratio. N = 30 was selected for the simulation. Increasing the number of rays in the diffuse components beyond 10 did not lead to a significant change in the multipath behavior (the resulting standard deviations were identical to within 1%). The Doppler for the diffuse component was random with a uniform distribution between 0 and 0.1 Hz. The delays were also random uniformly distributed. A time series for the duration of 1500 s was generated for each elevation angle (15,50 and 90 ). Fig. 4 shows a multipath simulation for the Galileo BOC(2, 2) signal with 8 MHz bandwidth and CMR = 6 db. The receiver model is a first order, noncoherent early late DLL with 1 Hz bandwidth and a second-order Costas PLL with 10 Hz bandwidth. The DLL is aided with the PLL (carrier aiding). The multi-ray model serves as input to the receiver model and the error in the DLL tracking was recorded. For multipath mitigation, a narrow correlator with 0.1 chip spacing was used. Note that AWG was turned off for these simulations and only the effect of the multipath is shown. The bias of m in Fig. 4 is caused by precorrelation filtering and does not impact the positioning as it appears on all lines of sight. The multipath error, however, is of course not identical on all lines-of-sight and thus does have an impact. The simulations show, that the mean value for the modified S-curve is about 54 cm and a standard deviation (1σ) of 1.6 m. 7. Acquisition performance Fig. 5 presents a simulation of the acquisition performance limit of the conventional noncoherent acquisition scheme, which is described by [2] Λ std = s μ 2 = R{s μ } 2 + I{s μ } 2. (27) The results of the conventional acquisition method are compared to a variant of the differential correlation method, which is described by N/2 Λ diff,4 = R s 2μ s2μ 1 = N/2 R{s 2μ } R{s 2μ 1 }+I{s 2μ } I{s 2μ 1 }. (28) s μ is the coherently integrated predetection result as defined in (7) in Section 4. Fig. 5 compares the performance of both algorithms when white Gaussian noise interference as defined in (7) and denoted by w μ, is present. A theoretical derivation of the probability density functions for crosscorrelation and the correlation peak is presented in [7]. The

7 A. Schmid et al. /Int. J. Electron. Commun. (AEÜ) 59 (2005) Fig. 4. Multipath error on L1 BOC(2, 2) for the modified S-curve in 8 MHz bandwidth for the unfavorable multipath conditions with CMR = 6 db. C/N 0 Acquisition [dbhz] Sensitivity in Presence of Thermal Noise P f = P d = 0.9 Standard Acquisition Algorithm Differential Acquisition Algorithm N=2 N=2 N=10 N=10 N=20 N=20 C/N 0 Acquisition [dbhz] Sensitivity in Presence of Thermal Noise P f = P d = 0.9 Standard Acquisition Algorithm N=30 N=30 N=40 N=40 N=100 N=100 Differential Acquisition Algorithm Coherent Integration Period T c [ms] Coherent Integration Period T c [ms] Fig. 5. Minimum required C/N 0 for acquisition in presence of thermal noise. The dashed lines correspond to the conventional noncoherent method, while the solid lines denote to the differential correlation algorithm. sensitivity of the algorithms is determined in Fig. 5 by calculating the minimally required carrier power to noise power spectral density ratio C/N 0 necessary to detect the signal with a probability of detection P d = 0.9 and a probability of false alarm P f = N denotes the noncoherent or differentially coherent integration number, respectively, and T c the coherent integration period. A substantial sensitivity improvement by means of differential correlation of up to2 db is observed with respect to the conventional noncoherent method. These simulations were also confirmed by implementing the algorithms in the Institute of Geodesy and Navigation PC-based Experimental Software Receiver (ipexsr) [8]. 8. Navigation software The presence of more than 50 satellites, GPS and Galileo, significantly enhances the navigation accuracy particularly in difficult surroundings like deep urban and moderate indoor areas. Especially in urban canyons, with a small angle

8 304 A. Schmid et al. /Int. J. Electron. Commun. (AEÜ) 59 (2005) Table 2. Position determination in case of degraded satellite availability Satellites in view Position determination 4 3D-positioning with receiver clock error estimation 3 3D-positioning without clock error estimation or 2D-positioning with receiver clock error estimation 2 2D-positioning without receiver clock estimation 1 1D-positioning prediction of user position based on previous trajectory of vision to the open sky, the additional Galileo satellites and therefore more visible satellites make satellite positioning feasible at all. Basically, due to the increase of available measurements by a factor k, the expected positioning accuracy raises by around k or more due to the higher measurement accuracy of the Galileo signals. The position fix then has to be calculated using a combination of Galileo and GPS C/A signals. Depending on the availability of the satellites, the navigation software can determine the receiver position by different modes, as presented in Table 2. When more than four satellites with pseudo-range and Doppler measurements are available, the full receiver state which contains position, velocity, clock bias, and clock drift can be determined. If less than four independent measurements are available, the number of unknowns in the receiver state has to be decreased in order to still get a distinct solution, as shown in Table 2. For example, a user position can be determined based on measurements to two satellites using a priori knowledge of the altitude (e.g. last known altitude) together with calculating the receiver clock error based of the latest estimated receiver clock parameters. Of course these assumptions decrease the positioning accuracy, but still provide the position information to the user. The user state determination is accomplished by Kalman filtering [9]. In addition to simple least mean-squared position solutions, the advantage of Kalman filtering is the prediction of the future user state by propagation of its current state to the next epoch. This prediction is used to estimate the Doppler search space for all satellites, which leads to shorter times to fix for reacquisition or single-shot measurements. Another advantage of the new navigation software is the computation of the receiver position based on spreading code phase measurements. Knowledge of only the code phase introduces an ambiguity that can be resolved by using an approximate user position. The single-shot positioning mode is developed for applications like mobile phones where minimal power consumption is an important requirement. It is planned to aid the positioning algorithm by additional information, such as the mobile phone cell ID. In addition, the navigation software offers a continuous positioning mode, developed for automotive applications. 9. Conclusion The paper has outlined a combined Galileo/GPS receiver concept to be developed during the first phase of the HIGAPS project. It is specially tailored for LBSs, E-911, and consumer market applications. A customized receiver architecture was presented in order to meet the extraordinary important requirements of low cost, low energy consumption, and enhanced sensitivity. The low- IF topology of the RF front-end circumvents the wellknown problems of zero-if topologies. The highly parallel digital baseband architecture utilizes the benefits of a low-power, high integration CMOS process. It enables single-shot measurements with short time to fix, high sensitivity, low power consumption, and low implementation complexity. Differential correlation provides sensitivity gain compared to conventional noncoherent integration. An adequate navigation processor supports single shot measurements and jointly utilizes Galileo and GPS. Acknowledgements The investigations and developments of this work are supported within the scope of the research project HIGAPS, which is jointly sponsored by the Bavarian Ministry of Economic Affairs, Infrastructure, Transport and Technology, as well as the German Aerospace Center (DLR). References [1] Crols J, Steyaert MSJ. Low-IF topologies for highperformance analog front ends of fully integrated receivers. IEEE Trans. Circuits Systems-II 1998;45(3): [2] Van Dierendonck AJ. GPS receivers. In: Parkinson BW, Spilker JJ, editors. Global positioning system: theory and applications, vol. 1. American Institute of Aeronautics and Astronautics, [3] Schmid A, Neubauer A. Performance evaluation of differential correlation for single shot measurement positioning. In: Proceedings of the Institute of Navigation GNSS September [4] Van Dierendonck AJ, Fenton P, Ford T. Theory and performance of narrow correlator spacing in a GPS receiver. Navigation 1992;39(3): [5] Van Dierendonck AJ, Braasch MS. Evaluation of GNSS receiver correlation processing techniques for multipath and noise mitigation. In: Proceedings of the institute of navigation national technical meeting January p [6] Mattos GPh. Multipath elimination for the low-cost consumer GPS. In: Proceedings of the Institute of Navigation GPS September p

9 A. Schmid et al. /Int. J. Electron. Commun. (AEÜ) 59 (2005) [7] Ávila-Rodríguez JA, Eissfeller B, Pany T. A theoretical analysis of acquisition algorithms for indoor positioning, ESA ESTEC NAVITEC 2004, December [8] Pany T, Eissfeller B, Hein GW. ipexsr: the first PC based software GNSS receiver completely developed in Europe. In: Proceedings of the European navigation conference GNSS May [9] Grover Brown R, Hwang P. Introduction to random signals and applied Kalman filtering. 3rd ed., New York: Wiley; [10] Hein GW. et al. Performance of Galileo L1 signal candidates. In: Proceedings of the European navigation conference GNSS May Andreas Schmid is a Concept Engineer at the Development Center NRW of Infineon Technologies AG, Germany, where he also pursues his Ph.D. degree. He received his Dipl.-Ing. degree in Computer Science and Communications Engineering from the University Duisburg-Essen, Germany in 2003, and M.Sc. degree in Communication Software and networks from the Nanyang Technological University Singapore in is the Director of the Institute of Electronics Engineering at the University of Erlangen-Nuremberg. Norbert Lemke is a Project Manager and Senior Systems Engineer for navigation applications and receiver development at IfEN GmbH. He holds a diploma (M.Sc.) in Aerospace Engineering from the Berlin University of Technology. Currently, he works in the field of navigation, integrity, receiver development and telematics. Günter Heinrichs received his Dipl.- Ing. degree in Communications Engineering from the University of Applied Science Aachen in 1988, a Dipl.-Ing. degree in Data Processing Engineering and a Dr.-Ing. degree in Electrical Engineering from the University Paderborn in 1991 and 1995, respectively. Since May 2002, he is the Head of business development and R&D management at IFEN GmbH. André Neubauer received his Dipl.- Ing. and Dr.-Ing. degrees in Electrical and Communications Engineering from the Gerhard-Mercator-University, Germany in 1992 and 1997, respectively. He is a Senior Principal for Concept Engineering at the Development Center NRW of Infineon Technologies AG, Germany. He is also an Associate Lecturer for Advanced Mobile Communications and Advanced Signal Theory at the University Duisburg-Essen. Henning Ehm received his diploma in Physics from the Technical University of Kaiserslautern, Germany, in He wrote his diploma thesis in the field of laser interferometry at IBM Germany. After his graduation Mr. Ehm joined the Institute of Electronics Engineering of the University of Erlangen-Nuremberg, Germany, as a Ph.D. student and Scientific Assistant, where he investigates analog front-end solutions for satellite navigation. Robert Weigel received his Dr.-Ing. and Dr.-Ing.habil. degrees in Electrical Engineering and Computer Science from the Munich University of Technology in Germany, in 1989 and 1992, respectively. From 1996 to 2002, he has been Director of the Institute for Communications and Information Engineering at the University of Linz. Since 2002, he Jón Winkel is the Head of Signals and Receivers at IfEN GmbH since He studied physics at the universities in Hamburg and Regensburg. From , he was a Research Associate at the Institute of Geodesy and Navigation at the Federal Armed Forces University. He received his Dr.-Ing. degree from the University of the Federal Armed Forces in Munich in Jose Angel Ávila-Rodríguez is a Research Associate and Ph.D. candidate at the Institute of Geodesy and Navigation at the University of the Federal Armed Forces Munich. He studied at the Technical Universities of Madrid, Spain, and Vienna, Austria, and has a M.S. in Electrical Engineering. His major areas of interest include among others the Galileo signal structure including codes, GNSS receiver design and performance. Roland Kaniuth received his diploma in Geodesy from the Technical University Munich. Since 2001, he is a Research Associate of the Institute of Geodesy and Navigation at the University FAF Munich. There, he is involved in different projects dealing with the development of GNSS software simulators and navigation algorithms.

10 306 A. Schmid et al. /Int. J. Electron. Commun. (AEÜ) 59 (2005) Thomas Pany has a Ph.D. in Geodesy from the Graz University of Technology and a M.S. in Physics from the Karl-Franzens University of Graz. Currently, he is a Research Associate at the Institute of Geodesy and Navigation at the University of Federal Armed Forces Munich. His major areas of interests include GPS/Galileo software receiver design, Galileo signal structure and GPS Meteorology. Günter Rohmer received his Master Degree in Electronic Engineering in 1988 and the Ph.D. in 1995 from the Technical University of Erlangen, Germany. Since 2001, he is Head of the Department at the Fraunhofer Institut for Integrated Circuits dealing with the development of components for satellite navigation receivers, indoor navigation and microwave localization systems. Bernd Eissfeller is a Full Professor and Vice-Director of the Institute of Geodesy and Navigation at the University of the Federal Armed Forces Munich. He is responsible for teaching and research in the field of Galileo/GPS/GLONASS and inertial technology. He obtained his Ph.D. in 1989 with a thesis on GPS/INS integration and his Habilitation with an analysis of GPS receivers. Matthias Overbeck received his Master Degree in Electronic Engineering in 1999 from the Technical University of Erlangen, Germany. Since 1999, he is a Research Associate at the Fraunhofer Institute for Integrated Circuits, responsible for the development of components for satellite navigation receivers, indoor navigation and microwave localization systems.