International Journal

on Marine Navigation

and Safety of Sea Transportation

Volume 4

Number 1

March 2010

1 INTRODUCTION

The Galileo system is a civil satellite navigation sys-

tem currently developed by the European Union.

The system will provide different types of services

for various missions, from a basic so called open

service (OS) to safety of life service (SoL) and pub-

lic regulated service (PRS). The Galileo signal struc-

ture is therefore very complicated and utilizes wide-

band signal with an AltBOC modulation for high

precision range measurement.

This paper is focused on the signal processing

methods of complete AltBOC Galileo E5 signal,

which is the most complex Galileo signal with the

widest frequency bandwidth. The optimal methods

of the GNSS signal processing are well known and

are based on the correlation reception. The naviga-

tion receiver calculates cross correlation function of

the received signal and a locally generated replica

signal and synchronized the replica on the received

signal.

The optimal correlator structure for Additive

White Gauss Noise Channel (AWGN) is the classi-

cal Early-Late correlator. The structure of such cor-

relator for GPS L1 C/A signal can be found in

Kaplan (1996). The replica generation for GPS L1

C/A signal is very simple, because the unknown

navigation data bits are modulated by the linear

BPSK modulation.

The proposed signal processing algorithm for the

GALILEO E5 signal is based on the same principles

like the GPS L1 C/A processing, but the correlator

structure is more complex.

2 GALILEO E5 SIGNAL CHARACTERISITC

The Galileo signals are defined in ESA (2008). The

cross correlation function of the signal was derived

in Kačmařík (2008) and can be written as follows

( )

[] [] j[]

Rm m m



= ℜ +ℑ





, (1)

where

( )

1234

11 22

1234

33 44

[] [] [] [].

[]

[][]

[] [] [] [].

[][]

sr sr sr sr

RnRnRnRn

m nSN m nSN

RnRnRnRn

m nSN m nSN

ρρ

∞

=−∞

∞

=−∞

+++



ℜ= +



−+ −







+++



−+ −





∑

(2)

and

( )

1234

[] [] [] [].

[] 2

[]

[] [] [] [].

[]

sr sr sr sr

RnRnRnRn

m nSN

RnRnRnRn

m nSN

∞

=−∞

∞

=−∞

 +−− 



ℑ= +



−







+−−





−





∑

. (3)

The replica signal of the Galileo E5 signal de-

pends on four secondary code bits and two naviga-

tion message bits. Since the secondary codes of the

data channels are multiplied by the navigation mes-

sage bits, the cross correlation function generally

depends on four bits. It results in sixteen possible

shapes of the cross correlation function between re-

ceived signal and generated replica. All of them are

depicted in Figure 1, where 1 or -1 in a chart title in-

dicates bits agreement or disagreement respectively.

Galileo AltBOC E5 Signal Characteristics for

Optimal Tracking Algorithms

F. Vejrazka, P. Kovar & P. Kacmarik

Czech Technical University in Prague, Czech Republic

ABSTRACT: The paper deals with an optimal processing of a Galileo E5 signal. A proposed correlator struc-

ture was developed on a base of a deep study of an E5 signal cross correlation function. Due to the non linear-

ity of the E5 AltBOC modulation the proposed correlator calculates the cross correlation function between a

received signal and a signal replica for all possible hypotheses of the navigation message bits. A correct peak

tracking verification is realized by implementation of a single side band correlator, which also serves for

course signal acquisition and secondary ranging code synchronization. The signal processing was verified on

the Galileo Giove A and Giove B satellites with very positive preliminary results.

The secondary code bits are known in the receiv-

er and can be generated after secondary codes syn-

chronization, but the navigation message bits remain

unknown.

Figure 1. Cross correlation function of received signal and replica for all combination of secondary codes bits

3 E5 CORRELATOR

The structure of the Early - Late correlator for Gali-

leo E5 signal is complicated since the complex Alt-

BOC modulation is used. There is no linear depend-

ency of the modulated signal on the navigation

message bits like in the GPS L1 C/A signal. The rep-

lica signal therefore must be generated for all possi-

ble hypotheses of the navigation unknown parame-

ters. In general case we have to generate 16 replicas.

The number of hypotheses can be reduced on four

hypotheses after the secondary code synchroniza-

tion.

This approach gives rise to the correlator struc-

ture shown in Figure 2. The first four correlator

branches serve for a calculation of the cross correla-

tion function for all combinations of the navigation

message bits. The fifth branch supports final signal

acquisition, secondary code synchronization and al-

so verification of the correct correlation peak track-

ing. This correlate branch processes only one signal

component of the Galileo E5 signal.

Figure 2. Galileo E5 correlator structure

4 CORRELATOR VERIFICATION

The proposed Galileo E5 correlator was verified by

the live Giove A and Giove B signals and on the

Galileo E5 signal generated by a signal generator.

This signal is processed by the GNSS software re-

ceiver EGR 2 which has been developed at the

Czech Technical University since 2000.

The block diagram of the receiver is drawn on

Figure 3. The Galileo signal is received by the wide-

band GNSS antenna equipped with the low noise

high dynamic range amplifier. The next receiver

block is a selective amplifier which splits partial

GNSS signals on L1 (E1), L2 and E5 (L5) frequen-

cies. The E5 signal is processed by a zero intermedi-

ate frequency receiver. The base band signal is digi-

talized and processed in Virtex 5 FPGA. The

measured data is sent via Ethernet to the PC work-

station for further processing.

Figure 3. EGR 2

Replica generator

Integrate and dump Hy-

pothesis 11

Replica generator

Integrate and dump Hy-

pothesis 01

Replica generator

Integrate and dump Hy-

pothesis 10

Replica generator

Integrate and dump Hy-

pothesis 00

EIh1

EQh1

LIh1

LQh1

Side band frequency

shift

Integrate and dump

SSB

TIC

Phase

EIh0

EQh0

LIh0

LQh0

EIh1

EQh1

LIh1

LQh1

EIh0

EQh0

LIh0

LQh0

Code

NCO Phase Ctr.

PRN-

Received

signal

Prn Ctr.

Baseband

Mixer

NCO

Code

PRN

Tab.

NCO Code Ctr.

Wideband

LNA

Selective

LNA

Zero IF re-

ceiver

ADC

Virtex 5

FPGA

PC work-

station

Antenna feeder

Ethernet

The proposed Galileo E5 correlator was devel-

oped in Matlab Simulink using with Xilinx System

Generator Toolbox. The correlator is controlled by

the embedded processor also integrated into the

FPGA.

5 EXPERIMENTAL RESULTS

This paragraph presents preliminary results of the

implemented Galileo E5 correlator. The results were

obtained by the receiver with not fully optimized

DLL and a PLL tracking loops. The signal was re-

ceived by the experimental antenna system equipped

with a helical RF filter. The noise figure of this an-

tenna is proximately 5 dB.

The plot of the pseudo range error and the carrier

phase error for the Galileo Giove A and Giove B

satellites are plotted on Figures 4 and 5. The stand-

ard deviation of these errors is in Table 1.

We are going to repeat these experiments with

higher performance GNSS antenna based on

PHEMT LNA with a noise figure 1 dB and a higher

performance selective LNA populated with the low

insertion loss and low distortion coaxial resonators

filters and with the fully optimized DLL and PLL

tracking loops. We believe that we will reach better

performance.

Figure 4. Code tracking error (

- Giove A,  – Giove B)

Table 1. Code and phase tracking error

______________________________________________

Satellite Giove A Giove B

______________________________________________

Standard deviation

of code tracking 0.202 0.204

error [m]

_____________________________________________

Standard deviation

of phase tracking 2.83 2.81

error [mm]

_____________________________________________

Figure 3. Phase tracking error (

- Giove A,  – Giove B)

6 CONCLUSIONS

This paper presents the Galileo AltBOC E5 signal

characteristics and on their base proposes the struc-

ture of the optimal Galileo E5 correlator for the

AWGN channel. The structure of the proposed cor-

relator is complicated due to the non linearity of the

AltBOC modulation of the navigation message bits

which requires to calculate the cross correlation

function between the received signal and the replica

for four hypotheses.

The developed correlator was tested in the soft-

ware receiver on the live Galileo signals with the

promising preliminary results. The final test is

planed to realize with the higher performance recep-

tion antenna and with the fully optimized PLL and

DLL tracking loops.

REFERENCES

ESA 2008. Galileo Open Service, Signal In Space Interface

Control Document, OS SIS ICD, Draft 1. European Space

Agency / European GNSS Supervisory Authority.

Kačmařík, P. & Kovář, P. & Vejražka, F. 2008. Galileo Alt-

BOC E5 signal characteristics for optimal tracking algo-

rithms. [CD ROM]. NAV08/ILA08. London: RIN.

Kaplan, D. 1996. Understanding GPS Principles and Applica-

tions. London: Artech House Inc.

Kovář, P. & Seidl, L. & Špaček, J. & Vejražka, F. 2006.

Software GNSS Receiver for Signal Experiments. In

IAIN/GNSS 2006. 12-th IAIN World Congress.

Proceedings, vol. 2. Jeju: Korean Institute of Navigation

and Port Research: 391-394.

0 1000 2000 3000 4000 5000 6000 7000 8000

-0.08

-0.06

-0.04

-0.02

0.02

0.04

0.06

time [ms]

Pseudorange error [PRN chips]

0 1000 2000 3000

4000 5000

6000 7000

8000

-0.05

0.05

time [ms]

Phase errot [cycles]