Advanced global navigation satellite system receiver design (797918), страница 22
Текст из файла (страница 22)
The Galileo signal at the DAC outputis centred at an Intermediate Frequency (IF) of 61.38 MHz, which is then upconverted and filtered to its respective signal bandwidth.IntermediateFrequencyoutputRF outputsE1 –1575 MHzAnalogueDevices ADCAD9755E5a – 1176 MHzE5b – 1207 MHzE6 – 1278 MHzRad tolerantActelRT54SX72SmodulatorFigure 7-3, MFUU for Giove-A148GNSS signal generators and the Giove-A satelliteThe combination of an FPGA and the AD9755 was suggested by SpirentCommunications, a leading supplier of GNSS signal simulations. The AD9755 hastwo input ports and alternately samples each port converting its value to an analoguecurrent.
If the sampling rate of the DAC is set to 4 times the desired signal IF, thenthe time difference between alternating samples is equivalent to 90 deg at the IF.Therefore, the in-phase (I) and quadrature (Q) modulation of the carrier can beachieved by feeding to the DAC synchronised but unique code sequences eachmodulated with an in-phase IF carrier.The modulation concept is simple and provides and elegant solution for GNSS signalgenerators. Let there be an analogue waveformu (t ) = A cos(ω 0 t + θ )7–1u (t ) = A cos(2πf 0 t + θ)7–2orwhich is then sampled at a rate fS = 4 f0 (carrier frequency). Without loss ofgenerality we write the corresponding data sequence from t = 0u[k ] = u (t k ) = A cos(2πf 0 t k + θ )7–3 kπ= A cos + θ 2If we look at the even numbered samples thenu[0], u[2] , u[4] ....= +Acos(θ), −Acos(θ), +Acos(θ) ....while odd numbered samples areu[1], u[3] , u[5] ....= −Asin(θ), +Asin(θ), −Asin(θ) ....Even numbered samples can be identified as sign-alternating samples of an I-channelmodulation.
Odd numbered samples can be identified as sign-alternating samples of aQ- channel modulation149GNSS signal generators and the Giove-A satelliteWe create a parallel input stream with a sampling rate in each of fSi = 122.76 MHz, sothatu[2k ] = (− 1)k mI [k ]7–4u[2k + 1] = −(− 1)k mQ [k ]which gives us a sequence+mI[0], −mQ[0], −mI[1], +mQ[1], +mI[2], −mQ[2], −mI[3], +mQ[3], +mI[4] ....etc.Spectrally and with f0 = 61.38 MHz then output sampling rate fSo= 245.52 MHz of theDAC supports a spectrum from dc to the Nyquist frequency fNo = 122.76 MHz.Figure 7-4 shows a depiction of a BOC modulation being applied in I, and a PSKmodulation in Q.QIf (MHz)fNi= 61.38fSi=122.76f (MHz)f0= 61.38fNo=122.76fSo=245.52Figure 7-4The input sampling rate in either channel fSi = 122.76 MHz.
supports a low passbandwidth up to an input Nyquist rate fNi = 61.38 MHz.A diagram of the code generation and IF modulation architecture of the MFUU isshown in Figure 7-5. In this case, the IF is 61.38 MHz and 122.76 MHz is fed to the150GNSS signal generators and the Giove-A satelliteDAC, which is internally doubled to the desired sampling rate of 245.52 MHz. Thegreat benefit of this scheme is of practical implementation. Creating I and Qcomponents digitally within the FPGA would require clocking the FPGA at245.52 MHz, which is impractical even with the fastest FPGA architectures.However, the AD9755 supports sampling rates up to 300 MHz and faster DACs areavailable. Using the DAC to introduce the I/Q modulation reduces the requiredFPGA clocking frequency by a factor of 2.
Fine amplitude control of the signal isachieved through adjusting the 8-bit level passed to the DAC. Even greater amplitudecontrol could be achieved if necessary as the AD9755 supports up to 14-bits of inputprecision.10.23 MHz8CodeGenerator8LatchDACamplitudescalingDACLatch8SubcarrierGenerator8IFGenerator90°AD9755PhaseDetector112FPGA122.76 MHzFigure 7-5, Simplified MFUU modulation architectureThe MFUU combines two Galileo navigation signals in a simple QPSK modulation,which can be represented as follows.SQPSK (t ) = PI × c1 (t ) × d1 (t ) × cos(ω Lt ) + PQ × c2 (t ) × d 2 (t ) × sin(ωLt )7–5PI and PQ are the in-phase and quadrature signal powers respectively, d(t) is thenavigational data applied to the signal and ω L is the L-band carrier frequency.
Ifa (t ) is the PRN coding sequence thenc(t ) = a (t ) × sgn [sin(2π f S t )]7–6for a BOC modulated signal with sub-carrier frequency fS and151GNSS signal generators and the Giove-A satellitec(t ) = a (t )7–7for PSK modulated signals. A spectrum analyser of the plot of the MFUU Galileo E1BOC(15, 2.5) signal is shown in Figure 7-6. The Galileo signals generated by theSSTL signal chain were verified and tested in the laboratory and in-orbit using theSeptentrio Galileo experimental test receiver (GETR) [De Wilde et al 2004] [Rooneyet al 2007].Figure 7-6, Spectrum analyser plot of the Galileo BOC(15, 2.5) –cosine signal from the MFUUThe MFUU on Giove-A will produce only QPSK signals and is based on Actel antifuse FPGAs, which are one-time programmable.
However, this architecture iscapable of producing all Galileo signals currently specified and representative GPSsignals for receiver testing. Therefore, a more flexible IF signal generator basedaround a re-programmable FPGA has been subsequently produced for this research.This IF signal generator (see Figure 7-7), reuses the modulation techniqueimplemented in the MFUU and can combine GNSS signals using QPSK, interplex orAltBOC modulations as desired for receiver testing.152GNSS signal generators and the Giove-A satelliteOven controlledoscillator (10.23MHz)FPGAmodulatorIF outputHigh speedDACFigure 7-7, Bench IF signal generatorAn equivalent to the interplex modulation also called modified hexaphase [Ries et al2002] intended for Galileo [ESA and GJU 2006], can be produced by the coherentadaptive subcarrier modulation (CASM) technique [Dafesh et al 2000].
Equation 7–5 shows a representation of the combination of two signals on a single carrier.Interplex modulation introduces a third signal, θ in the form of a phase modulation asfollows [Dafesh et al 1999].S INT (t ) = PI × c1 (t ) × cos(ωt + θ (t )) − PQ × c2 (t ) × sin(ωt + θ (t ))[− [P × c (t ) × sin(θ (t )) + P]× c (t ) × cos(θ (t ))]sin(ωt )7–8= PI × c1 (t ) × cos(θ (t )) − PQ × c2 (t ) × sin(θ (t )) cos(ωt )I1Q2If θ (t ) is another switching type signal written asθ (t ) = m × c 2 (t ) × c3 (t )7–9where m is the modulation index, which determines the relative distribution of thetransmitter power between the signals and c(t ) ∈ (+ 1,−1) .
The resulting modulation isequivalent to interplex modulation and can be written as follows.153GNSS signal generators and the Giove-A satellite PI × c2 (t ) × cos(m) PI × c1 (t ) × cos(m)S INT (t ) = cos(ωt ) + sin(ωt )− PQ × c3 (t ) × sin(m)+ PQ × c1 (t ) × c2 (t ) × c3 (t ) × sin(m)7–10The combination of all three PRN code sequence in the quadrature arm is called theinter-modulated product (IMP). It is generally the aim of the designer to minimisethe power contribution of the IMP as most receivers will not process this combination.The relative power levels of the interplex modulated GPS and Galileo signals areshown in Table 7-1.Table 7-1, Relative signal powers of interplex modulated GNSS signalsPRN codeEquivalentEquivalentRelative signal powerssequenceGalileo E1 orGPS L1 orGalileo E1 and E6GPSE6 signalL2 signal(m = 0.6155)(m = 0.66)c1AC/A or L2(C)22.2%22.2%c2BP(Y)44.4%40.2%c3CM22.2%24.2%c1 × c2 × c3--11.1%13.4%With regard to signal generation, the interplex modulation can be achieved simply byadjusting the I and Q levels of the DAC for each of the 8 possible states of the threeinput signals in accordance with the chosen modulation index.
Table 7-2 shows themapping of input signal state to 8-bit (0 to 255) DAC amplitude levels for GPS andGalileo signals1. In this representation the value of 127 should be regarded as thezero point. The I/Q plots of GPS and Galileo signal generated as a result of interplexmodulation are shown in Figure 7-8 and Figure 7-9 respectively. It can be seen thatthe IMP controls the amplitude of the quadrature contribution in order to maintainconstant envelope signal.154GNSS signal generators and the Giove-A satelliteTable 7-2 Example DAC amplitude mapping for interplex modulated GNSS signalsSignal stateDAC input levelGalileo E1 and E6GPS(m = 0.6155)(m = 0.66)c1c2c3c1 × c2× c3IQIQ1111-1-1-1-111-1-111-1-11-11-11-11-11-1-11-111-11272477127127247712725416916925408585012424912424951305130254161093161254930(c1 (t ) = +1, c2 (t ) = c3 (t ) ) c1 (t ) = +1, c2 (t ) = −1, c3 (t ) = +1 c1 (t ) = +1, c2 (t ) = +1, c3 (t ) = −1Q10.750.50.25I-1-0.75-0.5-0.250.250.50.751-0.25-0.5 c1 (t ) = −1, c2 (t ) = −1, c3 (t ) = +1-0.75-1 c1 (t ) = −1, c2 (t ) = +1, c3 (t ) = −1(c1 (t ) = −1, c2 (t ) = c3 (t ) )Figure 7-8, I Q plots for Galileo interplex modulated signals (m = 0.6155)155GNSS signal generators and the Giove-A satellite(c1 (t ) = +1, c2 (t ) = c3 (t ) ) c1 (t ) = +1, c2 (t ) = +1, c3 (t ) = −1Q c1 (t ) = +1, c2 (t ) = −1, c3 (t ) = +110.750.50.25-0.75-0.5-0.250.250.50.75I-0.25-0.5 c1 (t ) = −1, c2 (t ) = −1, c3 (t ) = +1-0.75-1 c1 (t ) = −1, c2 (t ) = +1, c3 (t ) = −1(c1 (t ) = −1, c2 (t ) = c3 (t ) )Figure 7-9, I Q plots for GPS interplex modulated signals (m = 0.66)The advantage of interplex modulation is that it is in theory a constant envelopemodulation.
![](https://s.studizba.com/z.php?f=/templates/si/i/noavatar.png&w=100&h=100&t=1"){kind=avatar}
