Title of Invention	"A METHOD AND DEVICE FOR DEMODULATING ALTERNATE BINARY OFFSET CARRIER SIGNALS
Abstract	A method and device for demodulating alternate binary offset carrier signals are disclosed. The method comprising at least two subcarriers (E5a, E5b) each having an in-phase and a quadrature component modulated by pseudo-random codes, the quadrature components being modulated by dataless pilot signals, the in-phase components being modulated by data signals, said method involving steps of: converting the alternate binary offset carrier signals into an intermediate frequency, band-pass filtering the converted signals and sampling the filtered signals, generating a carrier phase and carrier phase- rotating the sampled signals by said carrier phase, correlating the rotated sampled signals, and using the correlated rotated sampled signals as input of discriminators that sense carrier phase and code misalignments controlling local oscillators (4, 5), characterized in that it comprises steps of generating for each subcarrier (E5a, E5b) pseudo-random binary codes and a subcarrier phase, which are used to correlate the rotated sampled signals.

Title of Invention

"A METHOD AND DEVICE FOR DEMODULATING ALTERNATE BINARY OFFSET CARRIER SIGNALS

Abstract

A method and device for demodulating alternate binary offset carrier signals are disclosed. The method comprising at least two subcarriers (E5a, E5b) each having an in-phase and a quadrature component modulated by pseudo-random codes, the quadrature components being modulated by dataless pilot signals, the in-phase components being modulated by data signals, said method involving steps of: converting the alternate binary offset carrier signals into an intermediate frequency, band-pass filtering the converted signals and sampling the filtered signals, generating a carrier phase and carrier phase- rotating the sampled signals by said carrier phase, correlating the rotated sampled signals, and using the correlated rotated sampled signals as input of discriminators that sense carrier phase and code misalignments controlling local oscillators (4, 5), characterized in that it comprises steps of generating for each subcarrier (E5a, E5b) pseudo-random binary codes and a subcarrier phase, which are used to correlate the rotated sampled signals.

Full Text	FIELD OF THE INVENTION The present invention relates generally to Global Navigation Satellite System (GNSS) receivers and, in particular, to receivers that operate with Gaileo alternate binary offset carrier (AltBOC) satellite signals. BACKGROUND OF THE INVENTION Global navigation satellite system (GNSS) receivers, such as Galileo receivers, determine their global position based on signals received from orbiting G'^SS satellites. The GNSS satellites transmit signals using at least one carrier, -sach carrier being modulated by at least a binary pseudorandom (PRN) code, Which, consists of a seemingly random sequence of ones and zeros that is periodically repeated. The ones and zeros in the PRN code are referred to as "code chips" and the transitions in the bode from one to zero or zero to one, which occ jr at "code chip times" are referred to as "code chip transitions". Each GNSS satellite uses a unique PRN code, and thus, a GNSS receiver can associate a received signal with a particular satellite by determining which PRN code is included in the signal. ! The GNSS receiver calculates the difference between the time a satellite transmits its signal and the time that the receiver receives the signal. The receiver then calculates its distance, or "pseudo-range" from the satellite based on the associated time difference. Using the pseudo-ranges from at least four satellites, the receiver determines its global position. To determine the time difference, the GNSS receiver synchronizes a locally generated PRN code with the PRN code in the received signal by aligning the code chips in each of the codes. The GNSS receiver then determines how much the locally-generated PRN code is shifted in time from the known timing of the satellite PRN code at the time of transmission, and calculates the associated pseudo-range. The more closely the GNSS receiver aligns the locally-generated PRN code with the PRN code in the received signal, the more precisely the GNSS receiver can determine the associated time difference and pseudo-range and, in turn, its global position. The code synchronization operations include acquisition of the satellite PRN code and tracking the code. To acquire the PRN code, the GNSS receiver generally makes a series of correlation measurements that are separated in :ime by a code chip. After acquisition, the GNSS receiver tracks the received code. It generally makes "Early-Minus-Late" correlation measurements, ! i.e. measurements of the difference between (i) a correlation measurement i associated with the PRN code in the received signal and an early version of the locally-generated PRN code, and (ii) a correlation measurement associated with the PRN code in the received signal and a late version of the local PRN dode. The GNSS receiver then uses the early-minus-late measurements in a delay tlock loop (DLL), which produces an error signal that is proportional to\| the misalignment between the local and the received PRN codes. The error signal is used, in turn, to control the PRN code generator, which shifts the local PRN code essentially to minimize the DLL error signal. The GNSS receiver also typically aligns the satellite carrier with a local carrier using correlation measurements associated with a punctual version of the local PRN code. To do this the receiver uses a carrier tracking phase lock loop (PLL). The European Commission and the European Space Agency (ESAJ are developing a GNSS known as Galileo. Galileo satellites will transmit; two signals in the E5a band (1176.45 MHz) and two signals in the E5b band (1207.14 MHz) as a composite signal with-a center frequency of 1191.795 MHz and a bandwidth of at least 70 MHz, using a AltBOC modulation. \| The generation of the AltBOC signal is described in the document of the Galileo Signal Task Force of the European Commission "Status of Galileo Frequency and Signal Design", G. W. Hein, J. Godet, JX. Issler, J.C. Martin, P. Erhaid, R. Lucas-Rodriguez and T. Pratt, 25.09.2002, published at the following address: http://europa.eu.mt/comm/dgs/energy_transport/galileo/documents/teclmical_en .htm. Like the GPS satellites, the Galileo satellites each transmit unique PRN codes and a Galileo receiver can thus associate a received signal with a particular satellite. Accordingly, the Galileo receiver determines respective pseudo-ranges based on the difference between the times the satellites transmit the signals and then times the receiver receives the AltBOC signals. A standard linear offset carrier (LOC) modulates a time domain signal by a sine wave sin(a>0t), which shifts the frequency of the signal to both an up/per sideband and a corresponding lower sideband. The BOC modulation accomplishes the frequency shift using a square wave, or sign(sin(ffi0t)), and is generally denoted as BOC(fs, fc), where fs is the subcarrier (square wave) frequency and fc is the spreading code chipping rate. The factors of 1.023 ftJTHz are usually omitted from the notation for clarity so a BOC (15.345 MHz, 1(1).23 MHz) modulation is denoted BOC (15,10). j i j The modulation of a time domain signal by a complex exponential e"0' shifts the frequency of the signal to the upper sideband only. The goal of the AltSOC modulation is to generate in a coherent manner the E5a and E5b bands, which are respectively modulated by complex exponentials, or subcarriers, suchjthat the signals can be received as a wideband "BOC-like signal". The E5a and\|E5b bands each have associated in-phase (I) and quadrature (Q) spreading, or PKN, codes, with the E5a codes shifted to the lower sideband and the E5b cbdes shifted to the upper sideband. The respective E5a and E5b quadrature carriers are modulated by dataless pilot signals, and the respective in-phase carriers are modulated by both PKN codes and data signals. ■ I i The AltBOC modulation offers the advantage that the E5a (I and Q) and Ei5b (I and Q) signals can be processed independently as traditional BPSK(10) (Binary Phase-Shift Keying) signals, or together, leading to tremendous performances in terms of tracking noise and multipath. For the derivation of the demodulation principle of the AltBOC modulation, it is sufficient to approximate the base-band AltBOC signal by its AltLOC (Alternate Linear Offset Carrier) counterpart: where: - ci(t) is the PRN code of the E5b-data component (E5bl) and di(t) is the corresponding bit modulation; - c2(t) is the PKN code of the E5a-data component (E5al) and d2(t) is the corresponding bit modulation; j - c3(t) is the PKN code of the E5b-pilot component (E5bQ); - c4(t) is the PEN code of the E5a-pilot component (E5aQ); - the exponential factors represent the subcarrier modulation, of E5a and E5b; - cos is the side-band offset pulsation: cos = 27i;fs, with fs=15.345 MHz. In reality, s(t) contains additional product terms and the subcarrier exponentials are quantized. This effect will not be explicitly included in the equations for the sake of clarity. s(t) is modulated on the E5 carrier at 1191.795 MHz. Most previous publications present AltBOC from a satellite payload perspective, i.e. from a transmitter viewpoint. The receiver side processing has received very little attention so far. The publication "Comparison of AWGN Code Tracking Accuracy; for Alternative-BOC, Complex-LOC and Complex-BOC Modulation Options in Galileo E5-Band, M. Soellner and Ph. Erhard, GNSS 2003, April 2003, discloses the principle of a AltBOC receiver architecture for tracking the AltBOC pilot component, as shown in Fig. 1. In Fig. 1, the AltBOC receiver receives over an antenna 1 a signal that includes AltBOC composite codes transmitted by all of the satellites that are in view. The received signal is applied to a RF/IF stage 2 that, in a conventional maimer, converts the received signal RF to an intermediate frequency IF signal that has a frequency which is compatible with the otiher components of the receiver, falters the IF signal through a IF band-pass filter that has a band-pass at the desired center carrier frequency, and samples the filtered IF signal at a rate that satisfies the Nyquist theorem so as to produce corresponding digital in-phase (T) and quadrature (Q) signal samples in a known manner. The bandwidth of the'filter should be sufficiently wide to allow the primary harmonic of the AltBOC composite pilot code to pass, or approximately 51 MHz. The wide bandwidth results in relatively sharp code chip transitions in the received code, and!thus, s fairly well defined correlation peaks. j The AltBOC receiver comprises a local carrier oscillator 4, for example of the NCO type (Numerically Controlled Oscillator), synchronized with the IF frequency to generate a phase rotation angle on M bits which is applied to a phase rotator 3 receiving the IF signal samples on N bits. The phase rotated signal samples on N bits delivered by the phase rotator are applied to three complex correlators, each comprising a signal multiplier 10, 11, 12 and an integrator 13, 14, 15. The integrators sum the signal samples received during a predefined integration time Tint. The AltBOC receiver further comprises another local oscillator 5 of the TNICO type synchronized with the code chipping rate fc and which drives a complex AltBOC code generator 6 for locally generating complex PRN pilot codes for a given satellite. The generated pilot codes pass through a multi-hit delay line 7 comprising three cells E, P, L producing respectively early, prompt and Tate replicas of the local PRN codes which are applied respectively to an input of the multipliers 10,11, 12. The signals CE, CP and CL delivered by the integrators 13,14, 15 are then used to generate a carrier phase and code error signals which are used to drive; the NCO oscillators 4, 5. I The AltBOC code generator 6 presents the drawback of being complexj and multi-bit. Namely, it produces a quantized version of the Alt-LOC base-band signal (assuming only the pilot component is tracked) in the following form]: i Such a complex base-band signal is cumbersome to generate. The architecture shown in Fig. 1 also implies that all the operators (delay line, multipliers^ and integrators) operate on complex multi-bit numbers. SUMMARY OF THE INVENTION j An object of the present invention is to provide a simplified method and device for demodulating Galileo signals. i This object is achieved by a method for demodulating alternate binary offset carrier signals comprising at least two subcarriers each having an in-phase knd a quadrature component modulated by pseudo-random codes, the quadrature components being modulated by dataless pilot signals, the in-phase components being modulated by data signals, said method comprising steps of: - converting the alternate binary offset carrier signals into an intermediate frequency, band-pass filtering the converted signals and sampling the filtered signals, \| - generating a carrier phase and carrier phase-rotating the sampled signals by said carrier phase, and - correlating the rotated sampled signals. According to the invention, this method further comprises steps of generating for each subcarrier pseudo-random binary codes and a subcarrier phase, which are used to correlate the rotated sampled signals. j i According to a preferred embodiment of the invention, the method further comprises a step of translating said pseudo-random codes of said subcarriers into phase angles which are combined respectively with the subcarrier phases so as to obtain resultant phase angles for each subcarrier, said resultant phase angles being phase-shifted so as to obtain at least one early, a prompt and at least one late phase angles for each subcarrier, said correlation step comprising steps of phase-rotating said rotated sampled signals by said early, prompt and late phase angles of each subcarrier, for obtaining early, prompt and! late replicas of said sampled signals for each subcarrier, and integrating respectively the early, prompt and late replicas for each subcarrier during a predefined tiine. According to a preferred embodiment of the invention, the method further comprises a step of phase-rotating said rotated sampled signals by J said subcarriers phases' so as to obtain phase-rotated sampled signals for jeach subcarrier, before correlating said rotated sampled signals. According to a preferred embodiment of the invention, the method further comprises a step of bit-shifting said pseudo-random codes so as to obtain at least one early, a prompt and at least one late pseudo-random codes, said correlation step comprising steps of combining said phase-rotated sampled signals for each subcarrier with said early, prompt and late pseudo-random codes, and integrating the resulting signals during a predefined time, so obtain early, prompt and late correlation signals for each subcarrier,j said method further comprising a low speed post-correlation phase comprising !steps of phase-rotating the early and late correlation signals of each subcarrier respectively by opposite constant phase angles, and adding respectively the thus obtained early correlation signals of said subcarriers, the prompt correlation signals of said subcarriers and the thus obtained late correlation signals of said subcarriers so as to obtain respectively resultant early, prompt and late correlation signals. ; According to a preferred embodiment of the invention, the method further comprises a step of determining a combined carrier and subcarrier frequency for each subcarrier, the steps of phase-rotating by said carrier phase and said subcarriers phases being combined into a single phase rotation step for 6ach subcarrier using said combined carrier and subcarrier frequencies. j \| ■ According to a preferred embodiment of the invention, said correlation ^tep comprises steps of combining said phase-rotated sampled signals for each subcarrier respectively with the pseudo-random codes of said subcarrier, i and integrating during a predefined time the resulting signals for obtainitig a correlation signal for each subcarrier. ; According to a preferred embodiment of the invention, the method fufther comprises a low speed post-correlation phase comprising steps of combining the correlation signals for said subcarriers so as to obtain a prompt correlation signal used as an input of a PLL discrimination driving an oscillator controlling said carrier rotation step and a early-minus-late correlation signal used as an input of a DLL discrimination driving an oscillator controlling said code generation and said subcarrier phase generation. According to a preferred embodiment of the invention, the early-minus\|-late correlation signal is obtained from the correlation signals for said subcarriers E5a, E5b by the following formula: ; CE5,EmL ~ j(CE5a,0 - CE51),O)- According to a preferred embodiment of the invention, the DLL discrimination is of the type Dot-product power discrimination and performs the following operation: D = Real[CE5iEmL • CES;0 ], I where RealO is a function returning the real part of a complex number, the signal D being used to drive the oscillator controlling said code generation and said subcarrier phase generation. ! According to a preferred embodiment of the invention, the DLL discrimination performs the following operation: [ CB5,EmL — j (CsSa^O _ CE5b,o) ■ where ImagO is a function returning the imaginary part of a complex numb Jr. The invention also concerns a device for demodulating alternate binary offset carrier signals comprising at least two snbcarriers each having an in-phase ind a quadrature component modulated by pseudo-random codes, the quadrature components being modulated by dataless pilot signals, the in-phase components being modulated by data signals. According to the invention, this device comprises means for implementing the above-defined method. BRIEF DESCRIPTION OF THE ACCOMPANYING DRAWINGS The invention will be more clearly understood and other features and advantages of the invention will emerge from a reading of the following description given with reference to the appended drawings. Fig. 1 is a functional block diagram of a AltBOC demodulator channel according to prior art; ' Fig. 2 represents a curve of a single component correlation function for demodulating each component of a AltBOC signal; Fig. 3 is a Fresnel diagram of E5aQ and E5bQ single component correlation functions; Fig. 4 represents a curve of a combined correlation peak function combining E5aQ and E5bQ single component correlation functions; Fig. 5 is a functional block diagram of a AltBOC demodulator channel according to a first embodiment of the present invention; Fig. 6 is a functional block diagram of a AltBOC demodulator channel according to a second embodiment of the present invention; Fig. 7 is a functional block diagram of a AltBOC demodulator with two channels as shown in Fig. 6; i Fig. 8 is a functional block diagram of two channels of a AltBOC demodulator according to a third embodiment of the present invention; Fig. 9 is a Fresnel diagram of E5aQ and E5bQ single companent correlation functions obtained with the AltBOC demodulator of Fig. 9. Fig. 10 is a functional block diagram of a first embodiment of a receiver comprising the AltBOC demodulator of Fig. 8; j Fig. 11 is a functional block diagram of a second embodiment bf a receiver comprising the AltBOC demodulator of Fig. 8. Figs. 12 and 13 are functional block diagrams of a third and fourth embodiments of a AltBOC receiver derived from the receiver of 11. DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS ; The major characteristics of the invention will now be detailed. According to the AltBOC demodulation principle, the pilot channel is formed by! the combination of E5aQ and E5bQ signals. The AltBOC pilot signal is composed of the c3 and c4 components: where cos is the side-band offset pulsation: s = 2n:fs, with fs=15.345 MHz. In principle each component could be demodulated by correlating sP(t) with the code chip sequence, c;-sequence, multiplied by the complex conjugate of the corresponding subcarrier exponential, e.g. to track the c3(t) component^ the receiver must correlate with c3(t)-e_^tDs +n . The corresponding correlation function CE5bQ(t) can easily be derived (assuming an infinite bandwidth): ) where: j - the sign " - triangle(x) = ^ Mile. , [0 otherwise ; - x is the delay between the mcoming signal and the local code and subcarrier replicas; - Tint is the integration time; and - Tc is the chip length in units of time. The variations of the signal CE5bQ(x) as a function of the code tracking error are shown in Fig. 2. Curves 17 and 18 are respectively the real (I) and the imaginary (Q) components of this function, whereas curve 16 is the magnitude thereof. It can be seen that it is a complex function of T: if the local code and subcarrier replicas axe misaligned, energy moves from the I- to the Q-brahch. Such a correlation peak cannot be tracked as the code and carrier misalignrrients are not clearly separated: any code misalignment leads to a carrier phase tracking error. As the carrier loop is generally much faster than the code lotip, it will tend to zero the energy in the Q branch, resulting in the code loop seeing a pure BPSK correlation peak. j The additional information needed to make use of the BOC principle is the fact that the other side-band is coherently transmitted at a frequency distance of exactly 2fs = CHJTC. The CE5aQ(T) correlation function is given by correlating sP(t) withc4(t).e^t-'t/2): A Fresnel diagram as plotted in Fig. 3 provides an intuitive view of the complex QssaqC1) m^ CE5bQ(T) correlations. In this diagram, both correlations are represented as vectors in the I, Q plane. As the code delay T increases, CE5bQ and CE5aQ rotate with an angle +©sx and -CDST respectively, and their amplitude decreases according to the triangle function, leading to the two helixes as shown in the figure. A combined correlation peak function can be derived by summing the CE5a CE5bQ correlations, which corresponds to summing the vectors in Fig. 3: j As represented in Fig. 4, the function CE5Q(T) which corresponds to the AltBOC correlation peak function is real (curve 36) for all code delays, the imaginary part (curve 37) being null, and hence can be used for code tracking. ! I For the Pilot channel, the combined B5a/E5b correlation peak is simply the sum i of the individual E5a and E5b peaks. For the data channel, the same principle can be used, but the data bits have to be wiped off prior to the combination: the E5-Data correlation peak is given by: This bit estimation process makes the tracking channel less robust, especially at low signal to noise ratio (C/N0) where the probability of bit error is high. From this principle, five preferred embodiments of a AltBOC demodulator will be derived according to the invention. With a clever partitioning between pre- and post-correlation processing, the base-band processing of AltBOC can be done with little overhead with respect to traditional BPSK signals. The AltBOC demodulators presented below are derived assuming the Pilot channel is tracked, but the extension to the Data channel tracking is straightforward. It has been shown that building the AltBOC correlation peak involves correlating the incoming signal with c3(t)-e~j(c°st+7t/2) and c4(t)-ej(c°Bt"'t/2), and summing these two complex correlations. In the receiver, this is done in two identical channels, sharing the same local code and carrier oscillators. As explained above, demodulation of the c3 component involves correlating the mcoming signal with c3 (t) • e-J'ffist+,t '. This operation is equivalent to rotkting the incoming signal by an angle -mst - TC/2, followed by multiplying by the c3 PRN chips and integration. The multiplication by the code chips can be seen as an additional rotation by 0° if the chip is +1, or by 180° if the chip is -l.jThis observation leads to the first AltBOC demodulator channel architecture as represented in Fig. 5. In Fig. 5, the AltBOC demodulator channel receives over an antenna 1 a signal that includes the AltBOC composite codes transmitted by all of the satellites that are in view. The received signal is applied to a RF/IF stage 2 that converts the received signal RF to an intermediate frequency IF signal having a frequency which is compatible with the other components of the receiver, filters the IF signal through a IF band-pass filter that has a band-pass at the desired center carrier frequency, and samples the filtered IF signal at a rate that satisfies the Nyquist theorem so as to produce corresponding digital in-phase (Ij) and quadrature (Q) signal samples on N bits in a known manner. The bandwidth of the filter is sufficiently wide to allow the primary harmonic of the AltBOC composite pilot code to pass, or approximately 51 MHz. The wide bandwidth results in relatively sharp code chip transitions in the received code, and thus, fairly well defined correlation peaks. The AltBOC demodulator comprises a local oscillator 4, for example of the i NCO type (Numerically Controlled Oscillator), synchronized with! the frequency IF to generate a phase rotation angle on M bits which is applied! to a phase rotator 3 receiving the IF signal samples on N bits. The phase-rotated signal samples delivered by the phase rotator 3 are applied in parallel to three phase rotators 25, 26, 27 before being integrated in three respective integrators 28,29, 30 which sum their input signal samples during the integration time Tint. The AltBOC demodulator further comprises another local oscillator 5 of the NCO type synchronized with the code chipping rate fc and generating the jcode chipping rate and the subcarrier frequency fs = 1.5 fC3 for driving a subcarrier phase generator 20 and a E5b code generator 21. The output of E5b jcode generator 21 is connected to a PRN phase detector 22. The subcarrier phase generator 20 generates the phase of the subcarrier on M bits at the rate fs provided by the code NCO oscillator 5. The E5bQ code generator 21 generates the E5bQ code chips (0 or 1) at the rate fc given by the code NCO oscillator 5. The PRN phase detector translates the code chips (0 or 1) into a phase rotation angle 0 or TI. The respective output signals of the subcarrier phase generator and PRN phase detector are added by an adder 23, the output signal of the adder being a phase shift signal (real number coded on M bits) controlling a multi-bit delay like 24 with three cells E, P, L producing respectively early, prompt and late replichas of received PRN codes which are applied as phase shifts respectively to the phase rotators 25,26,27. The correlation signals Cnst,-!, CE5b,o and CE5b,i delivered by the integrators 28, 29, 30 are then used as input of discriminators that sense code and carrier phase rnisahgnments which are used to control the NCO oscillators 4, 5. The demodulator channel of Fig. 5 presents two main differences with respect to a traditional AltBOC demodulation channel as shown in Fig. 1: - the input to the delay-line 7 is a phase shift in the form of a real-valued signal; ! - the multiplication with the chip prior to the integration is replaced by a phase rotation. ; While the gate count required for this architecture is smaller than that of the standard architecture in Fig. 1, it is still large compared to the traditional 1-bit wide delay line. ! The architecture described in reference to Fig. 5 can be largely improved by noting that the E, P and L rotators 25, 26, 27 all rotate at the same frequency, but with a fixed phase difference. Namely, if the P rotator 26 applies a phase shift of-a>st - %I2, the E rotator 25 applies a phase shift of-ot)s(t+dTc/2) -j- u/2 and the L rotator 27 of-ffls(t-dTc/2) - n/2, where d is the Early-Late spacihg in units of chips, and Tc is the chip duration. This constant phase difference of ±cosdTc/2 can be taken out of the integration, and performed at low speed in post-correlation (after integration). This leads to the optimized architecture as presented in Fig. 6. Compared ip the architecture of Fig. 5: - each of the three rotators 25, 26, 27 is replaced by a respective Signal multiplier 33, 34, 35, : - a subcarrier rotator E5bQ 31 is inserted between the output of the carrier rotator 3 and the respective inputs of the signal multipliers 33, 34, 35;, and performs a phase rotation by e--1^3 ', - the multi-bit delay-line 24 is replaced by a one-bit wide code delay line 32 (the PRN phase detector being removed) and controlled directly by the E5b code generator 21, and - two signal multipliers 36, 37 respectively by e~Jct et e5" are inserted respectively at the output of the E and L integrators 28 and 30. The two signal multipliers 36, 37 belong to a low-speed post-correlation;stage (after integration), whereas the other part of this architecture belongs to ajhigh- speed pre-correlation stage. With this architecture, the only additional block with respect to a traditional BPSK demodulator is the subcarrier rotator 31, the phase of which is controlled by the code NCO oscillator 5. This architecture is mathematically equivalent to architecture of Fig. 5 if a is set to cosdTc/2. However, other values of a can be chosen to obtain virtually any other phase shift between the early audi late replicas. \| For clarity, the AltBOC demodulator architectures described in reference With Figs. 5 and 6 only show three complex correlators (early, punctual and late). In reality, detection of side-lobe tracking may require at least two additional correlators (very-early and very-late), but this is a straightforward extension of the structure. ! Thus the architecture represented in Fig. 5 or 6 can be extended to any nufnber of correlators. For instance, n early and m late correlators can be used, (each being feed with a respective cell of a delay line. CE5b)0 corresponds tb\| the prompt correlation. Typically, the early and late correlations are computed [with a delay of one cell with respect to the prompt correlation, i.e. they correspond to CE5b,i and CE5b,-i respectively. However, they can be set to any other delay. A typical application of the additional correlations is the detection of side peak tracking. Figs. 5 and 6 illustrate the architecture of one individual channel. In the AltBOC receiver, two of these channels for the E5 signal (one for E5a and one for E5b) are put together and the correlations are summed to produce an AltBOC correlation signal. Such a combined channel derived from1 the architecture of Fig. 6 is represented in Fig. 7. In Fig. 7, the architecture comprise a common RF/IF stage 2, carrier rotator 3, carrier NCO 4 and code NCO 5. i Each channel E5a, E5b comprises a subcarrier phase rotator 31a, 3 lib, a E5a/E5b code generator 21a, 21b feeding a respective delay line 32a, 32b,\|three respective correlators E, P, L, each including a signal multiplier 33a, 34a> 35a, 33b, 34b, 35b and an integrator 28a, 29a, 30a, 28b, 29b, 30b. The early and late branches of each channel E5a, E5b further comprise two respective signal multipliers 36a, 37a, 36b, 37b by a factor respectively equal to e_ja and e^. The subcarrier phase rotator 3 lb performs a phase rotation by e~^c°st+,t ', whereas the subcarrier phase rotator 3 la performs a phase rotation by eJ Channel E5a further comprises an additional signal multiplier 41a by a factor equal to -1, inserted between the code NCO 5 and the subcarrier rotator EJ5aQ 31a. The outputs of the two channels are added by three adders 42, 43,, 44 outputting respectively correlation signals CE5,i, CE5,o and CK-I. I Extending formulas (4) and (5), it can be derived that the C£5bik and &£Sa,k correlations are given by the following: ! where a = (BsdTc/2 = 2ixfsdTc/2. The Early-Late spacing d is determined by the clocking frequency of the delay line 32. Typically, d ranges from 0.1 to 1. ! For tracking, the receiver uses the CE5;k correlations to build code and carrier phase discriminators of which the output is proportional to the code and carrier phase tracking error respectively. The basis quantity used in the PLL discriminator is the punctual correlation CE5i0. The basic quantity used in the DLL discriminator is the difference between the Early and the Late correlations, also referred to as the Eaxly-niinus- Late correlation, and noted CE53mL. This difference reads: In the special case of d = l/(2fsTc) = 1/(215.345/10.23) = 1/3, a equals nil, and it can be shown that CE5iEmL is proportional to j(CE5a,o - CE5b]o) for Ismail tracking errors x. This fact leads to a dramatic reduction of the channel complexity, as only the punctual correlations (CE5a,o and CE5b,o) need to be I I I computed for both the code and carrier tracking. j This property can be demonstrated by reworking the expression for CE5)Emj, as follows, taking into account that formula (14) leads to: On the other hand, for small code tracking errors (T«1), j(CE5a,o-CE5b(o) is simply: J(CE5a,o -CE5bj0) = j(l-x)[e-ja^ -eJ'm°T] = 2sin(cosT) (18) This relation demonstrates that CE5,EmL is proportional to j(CE5ai0 - CE5b0). The factor (2 - d) is irrelevant as it is purely an amplification factor compensated for in the discriminator normalization. J This lead to an architecture as represented in Fig. 8, which is equivalent to the architecture of Fig 7 in the case of d = 1/3, though much simpler. With respect to the architectures of Figs. 6 and 7, this architecture does\| not comprises code delay lines 32a, 32b and have a single correlator for each! E5a and E5b codes. Each correlator comprises a single signal multiplier 51a, 51b receiving the output of the corresponding subcarrier rotator E5a and E5b 31a, 31b and the codes from the corresponding E5a and E5b code generator 21a] 21b and a single integrator 52a, 52b. The output signals CE5a]0 and CE5b)0 of the integrators 52a, 52b are applied to an adder 63 so as to obtain the punptual correlation signal CE5\|0, and to a comparator 64 and a multiplier by j 65 so jas to obtain the Early-minus-Late correlation signal CE5iEmL = j(CE5a,o - CE5b,o). It can be seen that this last architecture is extremely simple, as there is only one correlator needed per channel. Surprisingly, this leads to the conclusion that the AltBOC demodulator can be implemented very efficiently in terms of \ gate count, despites its apparent complexity. This last architecture shows that the tracking of the AltBOC signal can be done without any Early or Late correlator. This surprising result can be intuitiYely understood by drawing another Fresnel diagram, as in Fig 9. As established above, the code misalignment x is proportional to the angle q> between the CWo and the CES^O correlation vectors: cp = 2a>s%. It is also visible on the diagram that the vector j(CE5aio - CE5b,oX noted "E-L corr" in the diagram, obtained by subtracting the CE5b)0 vector from the CE5a>o vector, and by rotating the resulting vector by 90 degrees, is real, and has an amplitude proportional to the angle cp. This is the fundamental reason why the AltBOC code tracking does not heed Early and Late code replicas: the code misalignment can be derived solely from the punctual correlators. I Fig. 10. represents a receiver comprising the AltBOC demodulator of Fig. 8, and PLL (Phase-Lock Loop) and DLL (Delay-Lock Loop) controlling respectively the carrier NCO 4 and the Code NCO 5. The PLL comprises a discriminator 71 the output P of which is filtered by a PLL filter 72 before being applied to a control input of the carrier NCO 4.jThe PLL discriminator 71 is the arctan discriminator, which consists in computing the angle of the complex number CE5,o: P = Angle (CE5,O). \|(19) The DLL comprises a DLL discriminator receiving the correlation signal CE5,EmL and a DLL filter 76 connected to a control input of the code NCO 5i The DLL discriminator is of the type Dot-product power discriminator, which compute the signal D = Real(CE5;EmL-CE5>0). Thus the DLL discriminator comprises a complex conjugate function 73 to which the signal CEs,o is applied and a signal multiplier 74 for multiplying the signals provided by the multiplier by j 65 and the complex conjugate function 73. The signal D is then obtained by a function 75 extracting the real part of the complex signal delivered by the signal multiplier 74. After some algebraic manipulations, a simplified architecture as representjed in Fig. 11 can be derived from the architecture of Fig. 10, which requires fewer operations to compute the same DLL discriminator. ! According to the discrhriinator of Fig. 10: D = Real[CE5)EfflL-CE5(0] Thus, in Fig. 11, the DLL discriminator comprises a complex conjugate function 81 to which the correlation signal CE5B,O is applied and a signal multiplier 82 for multiplying the signal provided by complex conjugate function and the correlation signal CE5b\|0. The signal D is then obtained by a function ImagO 83 extracting the imaginary part of the complex signal delivered by; the signal multiplier 82. i A further modification of the architecture of Figure 11 would be; the replacement of the ImagO operator by an Angle() operator (i.e. a block providing the same functionality as the arctan discriminator 71). i The architecture of Figure 11 can be further optimized as shown in Fig. 12 by noticing that the phase rotation in the carrier rotator 3 followed by the phase rotation in the subcarriers rotators 31 a, 3 lb can be combined in one single phase rotation by a phase corresponding to the sum of the carrier and subcarrier phases. Thus in Fig. 12, the carrier rotator 3, the two subcarrier rotators 31a, 31b1 and the multiplier 41a of Fig. 11 are replaced with two phase rotators 92a and 92b (one for each channel E5a and E5b) receiving the down-converted signal ifrom the RF/TF stage 2. Besides, the subcarrier phase provided by the code NCO 4 is added by an adder 93a to the phase provided by the carrier NCO 3\ and subtracted therefrom by an adder 93b; the addition results being respectively applied to the phase rotators 92a, 92b of channels E5a, E5b. The architecture as shown in Fig. 13 can be derived from the previous architecture by replacing the Code NCO by a more simple NCO 95 delivering only the code chipping rate fc, and a frequency multiplier 96 by 1.5 applied to the code chipping rate fc so as to obtain the subcarrier frequency fs which is applied as input to the adders 93 a, 93b. This requires to duplicate the ctaier NCO 4, one for each channel E5a, E5b. The carrier frequency tracked b!y the PLL is applied to the adders 93a, 93b the respective outputs of which drive the carrier NCOs 91a, 91b of the two channels E5a5 E5b, so as to follow \| the respective combined carrier + subcarrier frequencies of the two channels E5a, E5b. j ■ I In this architecture the high-speed pre-correlation stages of E5a and JE5b channels remain identical. They both comprise a phase rotator 92a, 92b, jtwo NCOs 91a, 91b, a code generator 21a, 21b and a correlator. Moreover, if} the code NCO is duplicated so as to have one NCO per channel, each of the high- speed pre-correlation stages of E5a and E5b channels is identical to a traditional BPSK (Binary Phase-Shift Keying) channel, which offers great benefits id the design of a combined AltBOC/BPSK receiver. Of course, the optimizations performed in the architectures of Figs. 12 and 13 can be as well applied to the architectures of Figs. 5, 6 or 7. ! WE CLAIM: 1. A method performed by an electronic device for demodulating alternate binary offset carrier signals comprising at least two subcarriers (E5a, E5b) each having an in-phase component and a quadrature component, modulated by pseudo-random codes, the quadrature component of each subcarrier being modulated by dataless pilot signals, the in-phase component of each subcarrier being modulated by data signals, said method comprising steps of: converting the alternate binary offset carrier signals into an intermediate frequency, band-pass filtering the converted signals and sampling the filtered signals, generating a carrier phase and carrier phase-rotating the sampled signals by said carrier phase, correlating the rotated sampled signals,and using the correlated rotated sampled signals as input of discriminators that sense carrier phase and code misalignments controlling local oscillators (4, 5), characterized in that it comprises steps of generating for each subcarrier (E5a, E5b) pseudo-random binary codes and a subcarrier phase, which are used to correlate the rotated sampled signals. 2. The method as claimed in claim 1, comprising a step of translating said pseudo-random codes of said subcarriers into phase angles which are combined respectively with the subcarrier phases so as to obtain resultant phase angles for each subcarrier, said resultant phase angles being phase-shifted so as to obtain at least one early, a prompt and at least one late phase angles for each subcarrier, said correlation step comprising steps of phase-rotating said rotated sampled signals by said early, prompt and late phase angles of each subcarrier, for obtaining early, prompt and late replicas of said rotated sampled signals for each subcarrier, and integrating respectively the early, prompt and late replicas of said rotated sampled signals for each subcarrier during a predefined integration time. 3. The method as claimed in claim 1, comprising a step of phase-rotating said rotated sampled signals by said subcarriers phases so as to obtain phase- rotated sampled signals for each subcarrier (E5a, E5b), before correlating said rotated sampled signals. 4. The method as claimed in claim 3, comprising a step of bit-shifting said pseudo-random codes so as to obtain at least one early, a prompt and at least one late pseudo-random codes, said correlation step comprising steps of combining said phase-rotated sampled signals for each subcarrier with said early, prompt and late pseudo-random codes, and integrating the resulting signals during a predefined integration time, so as to obtain early, prompt and late correlation signals for each subcarrier (E5a, E5b), said method comprising a low speed post-correlation phase comprising steps of: phase-rotating the early correlation signals of each subcarrier respectively by opposite constant phase angles (ja, -ja), and adding the thus obtained early correlation signals of said subcarriers so as to obtain a resultant early correlation signal, phase-rotating the late correlation signals of each subcarrier respectively by said opposite constant phase angles, and adding the thus obtained late correlation signals of said subcarriers so as to obtain a resultant late correlation signal, adding the prompt correlation signals of said subcarriers and so as to obtain a resultant prompt correlation signal (CES,-L, CES.O, CE5,I). 5. The method as claimed in claim 3, comprising a step of determining a combined carrier and subcarrier phase for each subcarrier, the steps of phase- rotating by said carrier phase and the step of phase-rotating by said subcarriers phases being combined into a single phase rotation step for each subcarrier using said combined carrier and subcarrier phases. 6. The method as claimed in claim 3 or 5, wherein said correlation step comprises steps of combining said phase-rotated sampled signals for each subcarrier (E5a, E5b) respectively with the pseudo-random codes of said subcarrier, and integrating during a predefined integration time the resulting signals for obtaining a correlation signal (C^o, Creb.o) for each subcarrier. 7. The method as claimed in claim 6, comprising a low speed post- correlation phase comprising steps of combining through adders the correlation signals (CE5a,o, CKM) for said subcarriers (E5a, E5b) so as to obtain a resultant prompt correlation signal (CES.O) and a early-minus-late correlation signal, the prompt correlation signal being used as an input of a PLL discrimination driving a first oscillator (4) controlling said carrier rotation step, the early- minus-late correlation signal (CE5,EIIIL) being used as an input of a DLL discrimination driving a second oscillator (5) controlling said code generation and said subcarrier phase generation. 8. The method as claimed in claim 7, wherein the early-minus-late correlation signal (CESTUI.) is obtained from the correlation signals (CE5a,o, CE5b,o) for said subcarriers (E5a, E5b) by the following formula: CE5,EmL = j(CE5a,0 - CES^O) where CE5,EmL is the early-minus-late correlation signal, and CE5a,o and CE5b,o are the correlation signals for said subcarriers. 9. The method as claimed in anyone of claims 7 and 8, wherein the DLL discrimination is of the type Dot-product power discrimination and performs the following operation: D = Real[CE53nL-C^5o], where Real() is a function returning the real part of a complex number, CESEIHL is the early-minus-late correlation signal, and C^ is a complex conjugate of the resultant prompt correlation signal, the signal D being used to drive the second oscillator (5). 10. The method as claimed in any one of claims 7 and 8, wherein the DLL discrimination performs the following operation: D = Imag(CE5b,0-CE5a,o)- where Imag() is a function returning the imaginary part of a complex number, CE5b,ois the early-minus-late correlation signal and CE5a0is a complex conjugate of the prompt correlation signal, the signal D being used to drive the second oscillator (5). 11. A device for demodulating alternate binary offset carrier signals comprising at least two subcarriers (E5a, E5b) each having an in-phase and a quadrature component modulated by pseudo-random codes, the quadrature components being modulated by dataless pilot signals, the in-phase components being modulated by data signals, the device being configured to implement the method as claimed in any one of claims 1 to 10. Abstract A Method And Device For Demodulating Galileo Alternate Binary Offset Carrier Signals A method and device for demodulating alternate binary offset carrier signals are disclosed. The method comprising at least two subcarriers (E5a, E5b) each having an in-phase and a quadrature component modulated by pseudo-random codes, the quadrature components being modulated by dataless pilot signals, the in-phase components being modulated by data signals, said method involving steps of: converting the alternate binary offset carrier signals into an intermediate frequency, band-pass filtering the converted signals and sampling the filtered signals, generating a carrier phase and carrier phase- rotating the sampled signals by said carrier phase, correlating the rotated sampled signals, and using the correlated rotated sampled signals as input of discriminators that sense carrier phase and code misalignments controlling local oscillators (4, 5), characterized in that it comprises steps of generating for each subcarrier (E5a, E5b) pseudo-random binary codes and a subcarrier phase, which are used to correlate the rotated sampled signals.

Full Text

FIELD OF THE INVENTION
The present invention relates generally to Global Navigation Satellite System
(GNSS) receivers and, in particular, to receivers that operate with Gaileo
alternate binary offset carrier (AltBOC) satellite signals.
BACKGROUND OF THE INVENTION
Global navigation satellite system (GNSS) receivers, such as Galileo receivers,
determine their global position based on signals received from orbiting G'^SS
satellites. The GNSS satellites transmit signals using at least one carrier, -sach
carrier being modulated by at least a binary pseudorandom (PRN) code, Which,
consists of a seemingly random sequence of ones and zeros that is periodically
repeated. The ones and zeros in the PRN code are referred to as "code chips"
and the transitions in the bode from one to zero or zero to one, which occ jr at
"code chip times" are referred to as "code chip transitions". Each GNSS satellite
uses a unique PRN code, and thus, a GNSS receiver can associate a received
signal with a particular satellite by determining which PRN code is included in
the signal. !
The GNSS receiver calculates the difference between the time a satellite
transmits its signal and the time that the receiver receives the signal. The
receiver then calculates its distance, or "pseudo-range" from the satellite based
on the associated time difference. Using the pseudo-ranges from at least four
satellites, the receiver determines its global position.
To determine the time difference, the GNSS receiver synchronizes a locally
generated PRN code with the PRN code in the received signal by aligning the
code chips in each of the codes. The GNSS receiver then determines how much
the locally-generated PRN code is shifted in time from the known timing of the
satellite PRN code at the time of transmission, and calculates the associated
pseudo-range. The more closely the GNSS receiver aligns the locally-generated
PRN code with the PRN code in the received signal, the more precisely the
GNSS receiver can determine the associated time difference and pseudo-range

and, in turn, its global position.
The code synchronization operations include acquisition of the satellite PRN
code and tracking the code. To acquire the PRN code, the GNSS receiver
generally makes a series of correlation measurements that are separated in :ime
by a code chip. After acquisition, the GNSS receiver tracks the received code. It
generally makes "Early-Minus-Late" correlation measurements, ! i.e.
measurements of the difference between (i) a correlation measurement
i
associated with the PRN code in the received signal and an early version of the
locally-generated PRN code, and (ii) a correlation measurement associated with
the PRN code in the received signal and a late version of the local PRN dode.
The GNSS receiver then uses the early-minus-late measurements in a delay tlock
loop (DLL), which produces an error signal that is proportional to| the
misalignment between the local and the received PRN codes. The error signal is
used, in turn, to control the PRN code generator, which shifts the local PRN
code essentially to minimize the DLL error signal.
The GNSS receiver also typically aligns the satellite carrier with a local carrier
using correlation measurements associated with a punctual version of the local
PRN code. To do this the receiver uses a carrier tracking phase lock loop (PLL).
The European Commission and the European Space Agency (ESAJ are
developing a GNSS known as Galileo. Galileo satellites will transmit; two
signals in the E5a band (1176.45 MHz) and two signals in the E5b band
(1207.14 MHz) as a composite signal with-a center frequency of 1191.795 MHz
and a bandwidth of at least 70 MHz, using a AltBOC modulation. | The
generation of the AltBOC signal is described in the document of the Galileo
Signal Task Force of the European Commission "Status of Galileo Frequency
and Signal Design", G. W. Hein, J. Godet, JX. Issler, J.C. Martin, P. Erhaid, R.
Lucas-Rodriguez and T. Pratt, 25.09.2002, published at the following address:
http://europa.eu.mt/comm/dgs/energy_transport/galileo/documents/teclmical_en
.htm. Like the GPS satellites, the Galileo satellites each transmit unique PRN
codes and a Galileo receiver can thus associate a received signal with a
particular satellite. Accordingly, the Galileo receiver determines respective
pseudo-ranges based on the difference between the times the satellites transmit
the signals and then times the receiver receives the AltBOC signals.
A standard linear offset carrier (LOC) modulates a time domain signal by a sine

wave sin(a>0t), which shifts the frequency of the signal to both an up/per
sideband and a corresponding lower sideband. The BOC modulation
accomplishes the frequency shift using a square wave, or sign(sin(ffi0t)), and is
generally denoted as BOC(fs, fc), where fs is the subcarrier (square wave)
frequency and fc is the spreading code chipping rate. The factors of 1.023 ftJTHz
are usually omitted from the notation for clarity so a BOC (15.345 MHz, 1(1).23
MHz) modulation is denoted BOC (15,10). j
i
j
The modulation of a time domain signal by a complex exponential e*"0' shifts
the frequency of the signal to the upper sideband only. The goal of the AltSOC
modulation is to generate in a coherent manner the E5a and E5b bands, which
are respectively modulated by complex exponentials, or subcarriers, suchjthat
the signals can be received as a wideband "BOC-like signal". The E5a and|E5b
bands each have associated in-phase (I) and quadrature (Q) spreading, or PKN,
codes, with the E5a codes shifted to the lower sideband and the E5b cbdes
shifted to the upper sideband. The respective E5a and E5b quadrature carriers
are modulated by dataless pilot signals, and the respective in-phase carriers are
modulated by both PKN codes and data signals. ■
I
i
The AltBOC modulation offers the advantage that the E5a (I and Q) and Ei5b (I
and Q) signals can be processed independently as traditional BPSK(10) (Binary
Phase-Shift Keying) signals, or together, leading to tremendous performances in
terms of tracking noise and multipath.
For the derivation of the demodulation principle of the AltBOC modulation, it
is sufficient to approximate the base-band AltBOC signal by its AltLOC
(Alternate Linear Offset Carrier) counterpart:

where:
- ci(t) is the PRN code of the E5b-data component (E5bl) and di(t) is the
corresponding bit modulation;
- c2(t) is the PKN code of the E5a-data component (E5al) and d2(t) is the
corresponding bit modulation; j
- c3(t) is the PKN code of the E5b-pilot component (E5bQ);

- c4(t) is the PEN code of the E5a-pilot component (E5aQ);
- the exponential factors represent the subcarrier modulation, of E5a and E5b;
- cos is the side-band offset pulsation: cos = 27i;fs, with fs=15.345 MHz.
In reality, s(t) contains additional product terms and the subcarrier exponentials
are quantized. This effect will not be explicitly included in the equations for the
sake of clarity. s(t) is modulated on the E5 carrier at 1191.795 MHz.
Most previous publications present AltBOC from a satellite payload
perspective, i.e. from a transmitter viewpoint. The receiver side processing has
received very little attention so far.
The publication "Comparison of AWGN Code Tracking Accuracy; for
Alternative-BOC, Complex-LOC and Complex-BOC Modulation Options in
Galileo E5-Band, M. Soellner and Ph. Erhard, GNSS 2003, April 2003,
discloses the principle of a AltBOC receiver architecture for tracking the
AltBOC pilot component, as shown in Fig. 1.
In Fig. 1, the AltBOC receiver receives over an antenna 1 a signal that includes
AltBOC composite codes transmitted by all of the satellites that are in view.
The received signal is applied to a RF/IF stage 2 that, in a conventional maimer,
converts the received signal RF to an intermediate frequency IF signal that has a
frequency which is compatible with the otiher components of the receiver, falters
the IF signal through a IF band-pass filter that has a band-pass at the desired
center carrier frequency, and samples the filtered IF signal at a rate that satisfies
the Nyquist theorem so as to produce corresponding digital in-phase (T) and
quadrature (Q) signal samples in a known manner. The bandwidth of the'filter
should be sufficiently wide to allow the primary harmonic of the AltBOC
composite pilot code to pass, or approximately 51 MHz. The wide bandwidth
results in relatively sharp code chip transitions in the received code, and!thus,
s
fairly well defined correlation peaks. j
The AltBOC receiver comprises a local carrier oscillator 4, for example of the
NCO type (Numerically Controlled Oscillator), synchronized with the IF
frequency to generate a phase rotation angle on M bits which is applied to a
phase rotator 3 receiving the IF signal samples on N bits. The phase rotated
signal samples on N bits delivered by the phase rotator are applied to three
complex correlators, each comprising a signal multiplier 10, 11, 12 and an

integrator 13, 14, 15. The integrators sum the signal samples received during a
predefined integration time Tint.
The AltBOC receiver further comprises another local oscillator 5 of the TNICO
type synchronized with the code chipping rate fc and which drives a complex
AltBOC code generator 6 for locally generating complex PRN pilot codes for a
given satellite. The generated pilot codes pass through a multi-hit delay line 7
comprising three cells E, P, L producing respectively early, prompt and Tate
replicas of the local PRN codes which are applied respectively to an input of the
multipliers 10,11, 12.
The signals CE, CP and CL delivered by the integrators 13,14, 15 are then used
to generate a carrier phase and code error signals which are used to drive; the
NCO oscillators 4, 5.
I
The AltBOC code generator 6 presents the drawback of being complexj and
multi-bit. Namely, it produces a quantized version of the Alt-LOC base-band
signal (assuming only the pilot component is tracked) in the following form]:
i

Such a complex base-band signal is cumbersome to generate. The architecture
shown in Fig. 1 also implies that all the operators (delay line, multipliers^ and
integrators) operate on complex multi-bit numbers.
SUMMARY OF THE INVENTION j
An object of the present invention is to provide a simplified method and device
for demodulating Galileo signals. i
This object is achieved by a method for demodulating alternate binary offset
carrier signals comprising at least two subcarriers each having an in-phase knd a
quadrature component modulated by pseudo-random codes, the quadrature
components being modulated by dataless pilot signals, the in-phase components
being modulated by data signals, said method comprising steps of:
- converting the alternate binary offset carrier signals into an intermediate

frequency, band-pass filtering the converted signals and sampling the filtered
signals, |
- generating a carrier phase and carrier phase-rotating the sampled signals by
said carrier phase, and
- correlating the rotated sampled signals.
According to the invention, this method further comprises steps of generating
for each subcarrier pseudo-random binary codes and a subcarrier phase, which
are used to correlate the rotated sampled signals. j
i
According to a preferred embodiment of the invention, the method further
comprises a step of translating said pseudo-random codes of said subcarriers
into phase angles which are combined respectively with the subcarrier phases so
as to obtain resultant phase angles for each subcarrier, said resultant phase
angles being phase-shifted so as to obtain at least one early, a prompt and at
least one late phase angles for each subcarrier, said correlation step comprising
steps of phase-rotating said rotated sampled signals by said early, prompt and
late phase angles of each subcarrier, for obtaining early, prompt and! late
replicas of said sampled signals for each subcarrier, and integrating respectively
the early, prompt and late replicas for each subcarrier during a predefined tiine.
According to a preferred embodiment of the invention, the method further
comprises a step of phase-rotating said rotated sampled signals by J said
subcarriers phases' so as to obtain phase-rotated sampled signals for jeach
subcarrier, before correlating said rotated sampled signals.
According to a preferred embodiment of the invention, the method further
comprises a step of bit-shifting said pseudo-random codes so as to obtain at
least one early, a prompt and at least one late pseudo-random codes, said
correlation step comprising steps of combining said phase-rotated sampled
signals for each subcarrier with said early, prompt and late pseudo-random
codes, and integrating the resulting signals during a predefined time, so obtain early, prompt and late correlation signals for each subcarrier,j said
method further comprising a low speed post-correlation phase comprising !steps
of phase-rotating the early and late correlation signals of each subcarrier
respectively by opposite constant phase angles, and adding respectively the thus
obtained early correlation signals of said subcarriers, the prompt correlation
signals of said subcarriers and the thus obtained late correlation signals of said

subcarriers so as to obtain respectively resultant early, prompt and late
correlation signals. ;
According to a preferred embodiment of the invention, the method further
comprises a step of determining a combined carrier and subcarrier frequency for
each subcarrier, the steps of phase-rotating by said carrier phase and said
subcarriers phases being combined into a single phase rotation step for 6ach
subcarrier using said combined carrier and subcarrier frequencies. j
| ■
According to a preferred embodiment of the invention, said correlation ^tep
comprises steps of combining said phase-rotated sampled signals for each
subcarrier respectively with the pseudo-random codes of said subcarrier, i and
integrating during a predefined time the resulting signals for obtainitig a
correlation signal for each subcarrier. ;
According to a preferred embodiment of the invention, the method fufther
comprises a low speed post-correlation phase comprising steps of combining
the correlation signals for said subcarriers so as to obtain a prompt correlation
signal used as an input of a PLL discrimination driving an oscillator controlling
said carrier rotation step and a early-minus-late correlation signal used as an
input of a DLL discrimination driving an oscillator controlling said code
generation and said subcarrier phase generation.
According to a preferred embodiment of the invention, the early-minus|-late
correlation signal is obtained from the correlation signals for said subcarriers
E5a, E5b by the following formula: ;
CE5,EmL ~ j(CE5a,0 - CE51),O)-
According to a preferred embodiment of the invention, the DLL discrimination
is of the type Dot-product power discrimination and performs the following
operation: D = Real[CE5iEmL • CES;0 ], I
where RealO is a function returning the real part of a complex number,
the signal D being used to drive the oscillator controlling said code generation
and said subcarrier phase generation. !
According to a preferred embodiment of the invention, the DLL discrimination

performs the following operation: [
CB5,EmL — j (CsSa^O _ CE5b,o) ■
where ImagO is a function returning the imaginary part of a complex numb Jr.
The invention also concerns a device for demodulating alternate binary offset
carrier signals comprising at least two snbcarriers each having an in-phase ind a
quadrature component modulated by pseudo-random codes, the quadrature
components being modulated by dataless pilot signals, the in-phase components
being modulated by data signals. According to the invention, this device
comprises means for implementing the above-defined method.
BRIEF DESCRIPTION OF THE ACCOMPANYING DRAWINGS
The invention will be more clearly understood and other features and
advantages of the invention will emerge from a reading of the following
description given with reference to the appended drawings.
Fig. 1 is a functional block diagram of a AltBOC demodulator channel
according to prior art; '
Fig. 2 represents a curve of a single component correlation function for
demodulating each component of a AltBOC signal;
Fig. 3 is a Fresnel diagram of E5aQ and E5bQ single component
correlation functions;
Fig. 4 represents a curve of a combined correlation peak function
combining E5aQ and E5bQ single component correlation functions;
Fig. 5 is a functional block diagram of a AltBOC demodulator channel
according to a first embodiment of the present invention;
Fig. 6 is a functional block diagram of a AltBOC demodulator channel
according to a second embodiment of the present invention;
Fig. 7 is a functional block diagram of a AltBOC demodulator with two
channels as shown in Fig. 6;
i
Fig. 8 is a functional block diagram of two channels of a AltBOC
demodulator according to a third embodiment of the present invention;
Fig. 9 is a Fresnel diagram of E5aQ and E5bQ single companent
correlation functions obtained with the AltBOC demodulator of Fig. 9.
Fig. 10 is a functional block diagram of a first embodiment of a receiver
comprising the AltBOC demodulator of Fig. 8;

j
Fig. 11 is a functional block diagram of a second embodiment bf a
receiver comprising the AltBOC demodulator of Fig. 8.
Figs. 12 and 13 are functional block diagrams of a third and fourth
embodiments of a AltBOC receiver derived from the receiver of 11.
DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS ;
The major characteristics of the invention will now be detailed. According to
the AltBOC demodulation principle, the pilot channel is formed by! the
combination of E5aQ and E5bQ signals. The AltBOC pilot signal is composed
of the c3 and c4 components:

where cos is the side-band offset pulsation: s = 2n:fs, with fs=15.345 MHz.
In principle each component could be demodulated by correlating sP(t) with the
code chip sequence, c;-sequence, multiplied by the complex conjugate of the
corresponding subcarrier exponential, e.g. to track the c3(t) component^ the
receiver must correlate with c3(t)-e_^tDs +n . The corresponding correlation
function CE5bQ(t) can easily be derived (assuming an infinite bandwidth):
)
where: j
- the sign " - triangle(x) = ^ Mile. ,
[0 otherwise ;
- x is the delay between the mcoming signal and the local code and subcarrier
replicas;
- Tint is the integration time; and
- Tc is the chip length in units of time.
The variations of the signal CE5bQ(x) as a function of the code tracking error are
shown in Fig. 2. Curves 17 and 18 are respectively the real (I) and the
imaginary (Q) components of this function, whereas curve 16 is the magnitude

thereof. It can be seen that it is a complex function of T: if the local code and
subcarrier replicas axe misaligned, energy moves from the I- to the Q-brahch.
Such a correlation peak cannot be tracked as the code and carrier misalignrrients
are not clearly separated: any code misalignment leads to a carrier phase
tracking error. As the carrier loop is generally much faster than the code lotip, it
will tend to zero the energy in the Q branch, resulting in the code loop seeing a
pure BPSK correlation peak. j
The additional information needed to make use of the BOC principle is the fact
that the other side-band is coherently transmitted at a frequency distance of
exactly 2fs = CHJTC. The CE5aQ(T) correlation function is given by correlating sP(t)
withc4(t).e^t-'t/2):

A Fresnel diagram as plotted in Fig. 3 provides an intuitive view of the complex
QssaqC1) m^ CE5bQ(T) correlations. In this diagram, both correlations are
represented as vectors in the I, Q plane. As the code delay T increases, CE5bQ
and CE5aQ rotate with an angle +©sx and -CDST respectively, and their amplitude
decreases according to the triangle function, leading to the two helixes as shown
in the figure.
A combined correlation peak function can be derived by summing the CE5a CE5bQ correlations, which corresponds to summing the vectors in Fig. 3: j

As represented in Fig. 4, the function CE5Q(T) which corresponds to the AltBOC
correlation peak function is real (curve 36) for all code delays, the imaginary
part (curve 37) being null, and hence can be used for code tracking. !
I
For the Pilot channel, the combined B5a/E5b correlation peak is simply the sum
i
of the individual E5a and E5b peaks. For the data channel, the same principle
can be used, but the data bits have to be wiped off prior to the combination: the
E5-Data correlation peak is given by:

This bit estimation process makes the tracking channel less robust, especially at
low signal to noise ratio (C/N0) where the probability of bit error is high.
From this principle, five preferred embodiments of a AltBOC demodulator will
be derived according to the invention. With a clever partitioning between pre-
and post-correlation processing, the base-band processing of AltBOC can be
done with little overhead with respect to traditional BPSK signals.
The AltBOC demodulators presented below are derived assuming the Pilot
channel is tracked, but the extension to the Data channel tracking is
straightforward.
It has been shown that building the AltBOC correlation peak involves
correlating the incoming signal with c3(t)-e~j(c°st+7t/2) and c4(t)-ej(c°Bt"'t/2),
and summing these two complex correlations. In the receiver, this is done in
two identical channels, sharing the same local code and carrier oscillators.
As explained above, demodulation of the c3 component involves correlating the
mcoming signal with c3 (t) • e-J'ffist+,t '. This operation is equivalent to rotkting
the incoming signal by an angle -mst - TC/2, followed by multiplying by the c3
PRN chips and integration. The multiplication by the code chips can be seen as
an additional rotation by 0° if the chip is +1, or by 180° if the chip is -l.jThis
observation leads to the first AltBOC demodulator channel architecture as
represented in Fig. 5.
In Fig. 5, the AltBOC demodulator channel receives over an antenna 1 a signal
that includes the AltBOC composite codes transmitted by all of the satellites
that are in view. The received signal is applied to a RF/IF stage 2 that converts
the received signal RF to an intermediate frequency IF signal having a
frequency which is compatible with the other components of the receiver, filters
the IF signal through a IF band-pass filter that has a band-pass at the desired
center carrier frequency, and samples the filtered IF signal at a rate that satisfies
the Nyquist theorem so as to produce corresponding digital in-phase (Ij) and

quadrature (Q) signal samples on N bits in a known manner. The bandwidth of
the filter is sufficiently wide to allow the primary harmonic of the AltBOC
composite pilot code to pass, or approximately 51 MHz. The wide bandwidth
results in relatively sharp code chip transitions in the received code, and thus,
fairly well defined correlation peaks.
The AltBOC demodulator comprises a local oscillator 4, for example of the
i
NCO type (Numerically Controlled Oscillator), synchronized with! the
frequency IF to generate a phase rotation angle on M bits which is applied! to a
phase rotator 3 receiving the IF signal samples on N bits. The phase-rotated
signal samples delivered by the phase rotator 3 are applied in parallel to three
phase rotators 25, 26, 27 before being integrated in three respective integrators
28,29, 30 which sum their input signal samples during the integration time Tint.
The AltBOC demodulator further comprises another local oscillator 5 of the
NCO type synchronized with the code chipping rate fc and generating the jcode
chipping rate and the subcarrier frequency fs = 1.5 fC3 for driving a subcarrier
phase generator 20 and a E5b code generator 21. The output of E5b jcode
generator 21 is connected to a PRN phase detector 22. The subcarrier phase
generator 20 generates the phase of the subcarrier on M bits at the rate fs
provided by the code NCO oscillator 5. The E5bQ code generator 21 generates
the E5bQ code chips (0 or 1) at the rate fc given by the code NCO oscillator 5.
The PRN phase detector translates the code chips (0 or 1) into a phase rotation
angle 0 or TI.
The respective output signals of the subcarrier phase generator and PRN phase
detector are added by an adder 23, the output signal of the adder being a phase
shift signal (real number coded on M bits) controlling a multi-bit delay like 24
with three cells E, P, L producing respectively early, prompt and late replichas of
received PRN codes which are applied as phase shifts respectively to the phase
rotators 25,26,27.
The correlation signals Cnst,-!, CE5b,o and CE5b,i delivered by the integrators 28,
29, 30 are then used as input of discriminators that sense code and carrier phase
rnisahgnments which are used to control the NCO oscillators 4, 5.
The demodulator channel of Fig. 5 presents two main differences with respect
to a traditional AltBOC demodulation channel as shown in Fig. 1:

- the input to the delay-line 7 is a phase shift in the form of a real-valued
signal; !
- the multiplication with the chip prior to the integration is replaced by a phase
rotation. ;
While the gate count required for this architecture is smaller than that of the
standard architecture in Fig. 1, it is still large compared to the traditional 1-bit
wide delay line. !
The architecture described in reference to Fig. 5 can be largely improved by
noting that the E, P and L rotators 25, 26, 27 all rotate at the same frequency,
but with a fixed phase difference. Namely, if the P rotator 26 applies a phase
shift of-a>st - %I2, the E rotator 25 applies a phase shift of-ot)s(t+dTc/2) -j- u/2
and the L rotator 27 of-ffls(t-dTc/2) - n/2, where d is the Early-Late spacihg in
units of chips, and Tc is the chip duration. This constant phase difference of
±cosdTc/2 can be taken out of the integration, and performed at low speed in
post-correlation (after integration).
This leads to the optimized architecture as presented in Fig. 6. Compared ip the
architecture of Fig. 5:
- each of the three rotators 25, 26, 27 is replaced by a respective Signal
multiplier 33, 34, 35, :
- a subcarrier rotator E5bQ 31 is inserted between the output of the carrier
rotator 3 and the respective inputs of the signal multipliers 33, 34, 35;, and
performs a phase rotation by e--1^3 ',
- the multi-bit delay-line 24 is replaced by a one-bit wide code delay line 32
(the PRN phase detector being removed) and controlled directly by the E5b
code generator 21, and
- two signal multipliers 36, 37 respectively by e~Jct et e5" are inserted
respectively at the output of the E and L integrators 28 and 30.
The two signal multipliers 36, 37 belong to a low-speed post-correlation;stage
(after integration), whereas the other part of this architecture belongs to ajhigh-
speed pre-correlation stage.
With this architecture, the only additional block with respect to a traditional

BPSK demodulator is the subcarrier rotator 31, the phase of which is controlled
by the code NCO oscillator 5. This architecture is mathematically equivalent to
architecture of Fig. 5 if a is set to cosdTc/2. However, other values of a can be
chosen to obtain virtually any other phase shift between the early audi late
replicas. |
For clarity, the AltBOC demodulator architectures described in reference With
Figs. 5 and 6 only show three complex correlators (early, punctual and late). In
reality, detection of side-lobe tracking may require at least two additional
correlators (very-early and very-late), but this is a straightforward extension of
the structure. !
Thus the architecture represented in Fig. 5 or 6 can be extended to any nufnber
of correlators. For instance, n early and m late correlators can be used, (each
being feed with a respective cell of a delay line. CE5b)0 corresponds tb| the
prompt correlation. Typically, the early and late correlations are computed [with
a delay of one cell with respect to the prompt correlation, i.e. they correspond to
CE5b,i and CE5b,-i respectively. However, they can be set to any other delay. A
typical application of the additional correlations is the detection of side peak
tracking.
Figs. 5 and 6 illustrate the architecture of one individual channel. In the
AltBOC receiver, two of these channels for the E5 signal (one for E5a and one
for E5b) are put together and the correlations are summed to produce an
AltBOC correlation signal. Such a combined channel derived from1 the
architecture of Fig. 6 is represented in Fig. 7.
In Fig. 7, the architecture comprise a common RF/IF stage 2, carrier rotator 3,
carrier NCO 4 and code NCO 5.
i
Each channel E5a, E5b comprises a subcarrier phase rotator 31a, 3 lib, a
E5a/E5b code generator 21a, 21b feeding a respective delay line 32a, 32b,|three
respective correlators E, P, L, each including a signal multiplier 33a, 34a> 35a,
33b, 34b, 35b and an integrator 28a, 29a, 30a, 28b, 29b, 30b. The early and late
branches of each channel E5a, E5b further comprise two respective signal
multipliers 36a, 37a, 36b, 37b by a factor respectively equal to e_ja and e^. The
subcarrier phase rotator 3 lb performs a phase rotation by e~^c°st+,t ', whereas
the subcarrier phase rotator 3 la performs a phase rotation by eJ
Channel E5a further comprises an additional signal multiplier 41a by a factor
equal to -1, inserted between the code NCO 5 and the subcarrier rotator EJ5aQ
31a. The outputs of the two channels are added by three adders 42, 43,, 44
outputting respectively correlation signals CE5,i, CE5,o and CK-I.
I
Extending formulas (4) and (5), it can be derived that the C£5bik and &£Sa,k
correlations are given by the following: !

where a = (BsdTc/2 = 2ixfsdTc/2. The Early-Late spacing d is determined by the
clocking frequency of the delay line 32. Typically, d ranges from 0.1 to 1.
!
For tracking, the receiver uses the CE5;k correlations to build code and carrier
phase discriminators of which the output is proportional to the code and carrier
phase tracking error respectively.
The basis quantity used in the PLL discriminator is the punctual correlation
CE5i0. The basic quantity used in the DLL discriminator is the difference
between the Early and the Late correlations, also referred to as the Eaxly-niinus-
Late correlation, and noted CE53mL. This difference reads:

In the special case of d = l/(2fsTc) = 1/(2*15.345/10.23) = 1/3, a equals nil,
and it can be shown that CE5iEmL is proportional to j(CE5a,o - CE5b]o) for Ismail
tracking errors x. This fact leads to a dramatic reduction of the channel
complexity, as only the punctual correlations (CE5a,o and CE5b,o) need to be
I
I
I

computed for both the code and carrier tracking. j
This property can be demonstrated by reworking the expression for CE5)Emj, as
follows, taking into account that formula (14) leads to:

On the other hand, for small code tracking errors (T«1), j(CE5a,o-CE5b(o) is
simply:
J(CE5a,o -CE5bj0) = j(l-x)[e-ja^ -eJ'm°T] = 2sin(cosT) (18)
This relation demonstrates that CE5,EmL is proportional to j(CE5ai0 - CE5b0). The
factor (2 - d) is irrelevant as it is purely an amplification factor compensated for
in the discriminator normalization. J
This lead to an architecture as represented in Fig. 8, which is equivalent to the
architecture of Fig 7 in the case of d = 1/3, though much simpler.
With respect to the architectures of Figs. 6 and 7, this architecture does| not
comprises code delay lines 32a, 32b and have a single correlator for each! E5a
and E5b codes. Each correlator comprises a single signal multiplier 51a, 51b
receiving the output of the corresponding subcarrier rotator E5a and E5b 31a,
31b and the codes from the corresponding E5a and E5b code generator 21a] 21b
and a single integrator 52a, 52b. The output signals CE5a]0 and CE5b)0 of the
integrators 52a, 52b are applied to an adder 63 so as to obtain the punptual
correlation signal CE5|0, and to a comparator 64 and a multiplier by j 65 so jas to
obtain the Early-minus-Late correlation signal CE5iEmL = j(CE5a,o - CE5b,o). It can be seen that this last architecture is extremely simple, as there is only one
correlator needed per channel. Surprisingly, this leads to the conclusion that the
AltBOC demodulator can be implemented very efficiently in terms of \ gate
count, despites its apparent complexity.

This last architecture shows that the tracking of the AltBOC signal can be done
without any Early or Late correlator. This surprising result can be intuitiYely
understood by drawing another Fresnel diagram, as in Fig 9. As established
above, the code misalignment x is proportional to the angle q> between the CWo
and the CES^O correlation vectors: cp = 2a>s%. It is also visible on the diagram that
the vector j(CE5aio - CE5b,oX noted "E-L corr" in the diagram, obtained by
subtracting the CE5b)0 vector from the CE5a>o vector, and by rotating the resulting
vector by 90 degrees, is real, and has an amplitude proportional to the angle cp.
This is the fundamental reason why the AltBOC code tracking does not heed
Early and Late code replicas: the code misalignment can be derived solely from
the punctual correlators. I
Fig. 10. represents a receiver comprising the AltBOC demodulator of Fig. 8,
and PLL (Phase-Lock Loop) and DLL (Delay-Lock Loop) controlling
respectively the carrier NCO 4 and the Code NCO 5.
The PLL comprises a discriminator 71 the output P of which is filtered by a
PLL filter 72 before being applied to a control input of the carrier NCO 4.jThe
PLL discriminator 71 is the arctan discriminator, which consists in computing
the angle of the complex number CE5,o:
P = Angle (CE5,O). |(19)
The DLL comprises a DLL discriminator receiving the correlation signal
CE5,EmL and a DLL filter 76 connected to a control input of the code NCO 5i
The DLL discriminator is of the type Dot-product power discriminator, which
compute the signal D = Real(CE5;EmL-CE5>0). Thus the DLL discriminator
comprises a complex conjugate function 73 to which the signal CEs,o is applied
and a signal multiplier 74 for multiplying the signals provided by the multiplier
by j 65 and the complex conjugate function 73. The signal D is then obtained by
a function 75 extracting the real part of the complex signal delivered by the
signal multiplier 74.
After some algebraic manipulations, a simplified architecture as representjed in
Fig. 11 can be derived from the architecture of Fig. 10, which requires fewer
operations to compute the same DLL discriminator. !
According to the discrhriinator of Fig. 10:
D = Real[CE5)EfflL-CE5(0]

Thus, in Fig. 11, the DLL discriminator comprises a complex conjugate
function 81 to which the correlation signal CE5B,O is applied and a signal
multiplier 82 for multiplying the signal provided by complex conjugate function
and the correlation signal CE5b|0. The signal D is then obtained by a function
ImagO 83 extracting the imaginary part of the complex signal delivered by; the
signal multiplier 82. i
A further modification of the architecture of Figure 11 would be; the
replacement of the ImagO operator by an Angle() operator (i.e. a block
providing the same functionality as the arctan discriminator 71).
i
The architecture of Figure 11 can be further optimized as shown in Fig. 12 by
noticing that the phase rotation in the carrier rotator 3 followed by the phase
rotation in the subcarriers rotators 31 a, 3 lb can be combined in one single phase
rotation by a phase corresponding to the sum of the carrier and subcarrier
phases.
Thus in Fig. 12, the carrier rotator 3, the two subcarrier rotators 31a, 31b1 and
the multiplier 41a of Fig. 11 are replaced with two phase rotators 92a and 92b
(one for each channel E5a and E5b) receiving the down-converted signal ifrom
the RF/TF stage 2. Besides, the subcarrier phase provided by the code NCO 4 is
added by an adder 93a to the phase provided by the carrier NCO 3\ and
subtracted therefrom by an adder 93b; the addition results being respectively
applied to the phase rotators 92a, 92b of channels E5a, E5b.
The architecture as shown in Fig. 13 can be derived from the previous
architecture by replacing the Code NCO by a more simple NCO 95 delivering
only the code chipping rate fc, and a frequency multiplier 96 by 1.5 applied to
the code chipping rate fc so as to obtain the subcarrier frequency fs which is
applied as input to the adders 93 a, 93b. This requires to duplicate the ctaier
NCO 4, one for each channel E5a, E5b. The carrier frequency tracked b!y the
PLL is applied to the adders 93a, 93b the respective outputs of which drive the

carrier NCOs 91a, 91b of the two channels E5a5 E5b, so as to follow | the
respective combined carrier + subcarrier frequencies of the two channels E5a,
E5b. j
■
I
In this architecture the high-speed pre-correlation stages of E5a and JE5b
channels remain identical. They both comprise a phase rotator 92a, 92b, jtwo
NCOs 91a, 91b, a code generator 21a, 21b and a correlator. Moreover, if} the
code NCO is duplicated so as to have one NCO per channel, each of the high-
speed pre-correlation stages of E5a and E5b channels is identical to a traditional
BPSK (Binary Phase-Shift Keying) channel, which offers great benefits id the
design of a combined AltBOC/BPSK receiver. Of course, the optimizations performed in the architectures of Figs. 12 and 13
can be as well applied to the architectures of Figs. 5, 6 or 7. !

WE CLAIM:
1. A method performed by an electronic device for demodulating alternate
binary offset carrier signals comprising at least two subcarriers (E5a, E5b) each
having an in-phase component and a quadrature component, modulated by
pseudo-random codes, the quadrature component of each subcarrier being
modulated by dataless pilot signals, the in-phase component of each subcarrier
being modulated by data signals, said method comprising steps of:
converting the alternate binary offset carrier signals into an intermediate
frequency, band-pass filtering the converted signals and sampling the filtered
signals,
generating a carrier phase and carrier phase-rotating the sampled signals
by said carrier phase,
correlating the rotated sampled signals,and
using the correlated rotated sampled signals as input of discriminators
that sense carrier phase and code misalignments controlling local oscillators (4,
5),
characterized in that it comprises steps of generating for each subcarrier (E5a,
E5b) pseudo-random binary codes and a subcarrier phase, which are used to
correlate the rotated sampled signals.
2. The method as claimed in claim 1, comprising a step of translating said
pseudo-random codes of said subcarriers into phase angles which are combined
respectively with the subcarrier phases so as to obtain resultant phase angles for
each subcarrier, said resultant phase angles being phase-shifted so as to obtain
at least one early, a prompt and at least one late phase angles for each
subcarrier, said correlation step comprising steps of phase-rotating said rotated
sampled signals by said early, prompt and late phase angles of each subcarrier,
for obtaining early, prompt and late replicas of said rotated sampled signals for

each subcarrier, and integrating respectively the early, prompt and late replicas
of said rotated sampled signals for each subcarrier during a predefined
integration time.
3. The method as claimed in claim 1, comprising a step of phase-rotating
said rotated sampled signals by said subcarriers phases so as to obtain phase-
rotated sampled signals for each subcarrier (E5a, E5b), before correlating said
rotated sampled signals.
4. The method as claimed in claim 3, comprising a step of bit-shifting said
pseudo-random codes so as to obtain at least one early, a prompt and at least
one late pseudo-random codes, said correlation step comprising steps of
combining said phase-rotated sampled signals for each subcarrier with said
early, prompt and late pseudo-random codes, and integrating the resulting
signals during a predefined integration time, so as to obtain early, prompt and
late correlation signals for each
subcarrier (E5a, E5b), said method comprising a low speed post-correlation
phase comprising steps of:
phase-rotating the early correlation signals of each subcarrier
respectively by opposite constant phase angles (ja, -ja), and adding the thus
obtained early correlation signals of said subcarriers so as to obtain a resultant
early correlation signal,
phase-rotating the late correlation signals of each subcarrier respectively
by said opposite constant phase angles, and adding the thus obtained late
correlation signals of said subcarriers so as to obtain a resultant late correlation
signal,
adding the prompt correlation signals of said subcarriers and so as to
obtain a resultant prompt correlation signal (CES,-L, CES.O, CE5,I).

5. The method as claimed in claim 3, comprising a step of determining a
combined carrier and subcarrier phase for each subcarrier, the steps of phase-
rotating by said carrier phase and the step of phase-rotating by said subcarriers
phases being combined into a single phase rotation step for each subcarrier
using said combined carrier and subcarrier phases.
6. The method as claimed in claim 3 or 5, wherein said correlation step
comprises steps of combining said phase-rotated sampled signals for each
subcarrier (E5a, E5b) respectively with the pseudo-random codes of said
subcarrier, and integrating during a predefined integration time the resulting
signals for obtaining a correlation signal (C^o, Creb.o) for each subcarrier.
7. The method as claimed in claim 6, comprising a low speed post-
correlation phase comprising steps of combining through adders the correlation
signals (CE5a,o, CKM) for said subcarriers (E5a, E5b) so as to obtain a resultant
prompt correlation signal (CES.O) and a early-minus-late correlation signal, the
prompt correlation signal being used as an input of a PLL discrimination
driving a first oscillator (4) controlling said carrier rotation step, the early-
minus-late correlation signal (CE5,EIIIL) being used as an input of a DLL
discrimination driving a second oscillator (5) controlling said code generation
and said subcarrier phase generation.
8. The method as claimed in claim 7, wherein the early-minus-late
correlation signal (CESTUI.) is obtained from the correlation signals (CE5a,o, CE5b,o)
for said subcarriers (E5a, E5b) by the following formula:
CE5,EmL = j(CE5a,0 - CES^O)
where CE5,EmL is the early-minus-late correlation signal, and CE5a,o and
CE5b,o are the correlation signals for said subcarriers.

9. The method as claimed in anyone of claims 7 and 8, wherein the DLL
discrimination is of the type Dot-product power discrimination and performs
the following operation:
D = Real[CE53nL-C^5o],
where Real() is a function returning the real part of a complex number, CESEIHL
is the early-minus-late correlation signal, and C^ is a complex conjugate of
the resultant prompt correlation signal, the signal D being used to drive the
second oscillator (5).
10. The method as claimed in any one of claims 7 and 8, wherein the DLL
discrimination performs the following operation:
D = Imag(CE5b,0-CE5a,o)-
where Imag() is a function returning the imaginary part of a complex number,
CE5b,ois the early-minus-late correlation signal and CE5a0is a complex
conjugate of the prompt correlation signal, the signal D being used to drive the
second oscillator (5).
11. A device for demodulating alternate binary offset carrier signals
comprising at least two subcarriers (E5a, E5b) each having an in-phase and a
quadrature component modulated by pseudo-random codes, the quadrature
components being modulated by dataless pilot signals, the in-phase
components being modulated by data signals, the device being configured to
implement the method as claimed in any one of claims 1 to 10.

Abstract

A Method And Device For Demodulating Galileo
Alternate Binary Offset Carrier Signals
A method and device for demodulating alternate binary offset carrier
signals are disclosed. The method comprising at least two subcarriers (E5a,
E5b) each having an in-phase and a quadrature component modulated by
pseudo-random codes, the quadrature components being modulated by dataless
pilot signals, the in-phase components being modulated by data signals, said
method involving steps of: converting the alternate binary offset carrier signals
into an intermediate frequency, band-pass filtering the converted signals and
sampling the filtered signals, generating a carrier phase and carrier phase-
rotating the sampled signals by said carrier phase, correlating the rotated
sampled signals, and using the correlated rotated sampled signals as input of
discriminators that sense carrier phase and code misalignments controlling
local oscillators (4, 5), characterized in that it comprises steps of generating for
each subcarrier (E5a, E5b) pseudo-random binary codes and a subcarrier phase,
which are used to correlate the rotated sampled signals.

"A METHOD AND DEVICE FOR DEMODULATING ALTERNATE BINARY OFFSET CARRIER SIGNALS

Documents:

Inventors:

PCT Conventions: