EP1456686A1

EP1456686A1 - Method of processing a navigation signal containing data

Info

Publication number: EP1456686A1
Application number: EP02805387A
Authority: EP
Inventors: Nicolas Thales Intellectual Property Martin; Blandine Thales Intellectual Property Coatantiec
Original assignee: Thales SA
Current assignee: Thales SA
Priority date: 2001-12-18
Filing date: 2002-12-13
Publication date: 2004-09-15
Also published as: FR2833714A1; US20040260456A1; FR2833714B1; CA2470375A1; WO2003054575A1

Abstract

The invention relates to methods of processing satellite navigation signals which also contain data. The inventive method consists in using two nested carrier loops; one comprising a DBPSK differential-type extended demodulator (115) which is used to maintain the dynamics, and another comprising a BPSK-type non-extended demodulator (117), so as not to correlate the code errors. The invention enables the use of a much higher data transmission speed which requires the use of a Viterbi-type algorithm.

Description

Method for processing a navigation signal containing data.

The present invention relates to methods for processing navigation signals when they also include one or more additional modulations making it possible to transmit data which complement the information intrinsically contained in the elementary signal. It applies more particularly to the processing of navigation signals transmitted by satellites.

Radio navigation by satellite currently mainly uses the signals emitted by the GPS system, and incidentally those of the GLONASS system. In the future, it will use the signals from the GALILEO system. This radio navigation makes it possible to obtain the position of the receiver by processing the signals transmitted by at least three satellites and by performing phase measurements and code measurements on these signals. Code measurements are not very precise but unambiguous, while phase measurements are much more precise but ambiguous.

The signals transmitted by the satellites contain data bits which make it possible to transmit navigation messages. These messages are used in the calculation algorithms to determine the position, speed and time of the receiver.

In the systems currently used, of the GPS type, a data rate of the order of 50 Bauds is used, this allows on the one hand to extract without too much difficulty these data from the signal received, whose level is very low, and on the other hand not to disturb the main navigation signal. Furthermore, given the optimization thus chosen, it is not necessary to use an error correcting system and consequently each bit transmitted corresponds to a useful bit.

In the future systems currently under study, of the GALILEO type, we would like to transmit a lot more data, which implies using a digital bit rate of around 1000 Bauds.

Under these conditions the error rate increases considerably, which requires the use of an error correction code. It is therefore planned to use a system of the so-called "Forward Error Corrector" type, comprising a level 2 redundancy and a level 7 convolution. The error correction would then be done with a Viterbi type algorithm, for example, knows that to function properly it requires that the errors be uncorrelated. To demodulate the received signal so that the data rate does not disturb the demodulation of the navigation signal too much, we are led to use a differential discriminator of the type known as DBPSK. Such a discriminator uses to demodulate the information corresponding to one bit the information corresponding to the demodulation of the previous bit, hence the qualifier of "differential".

Under these conditions, the received phase rotates and it is then necessary to use an integration system, which makes it possible to obtain a reactive control over this phase.

Furthermore, in a differential demodulator of the DBPSK type, the difference in signs from one sample to the next is used to determine the value of the data bit received. As the signal is noisy, this means that the error on one bit depends on the value of 2 successive samples. Under these conditions the error is no longer uncorrelated and the Viterbi algorithm does not work. In order to nevertheless be able to obtain the data with a sufficiently low error rate, while keeping correct reception of the navigation signals, the invention proposes a method of processing a navigation signal containing data, in which a first DBPSK type carrier band for demodulating navigation signals, and in which a second BPSK type loop is also used to identify the data tests.

Other features and advantages of the invention will appear clearly in the following description, presented with reference to the appended figures which represent: - Figure 1, an overall diagram of the architecture of the loops used in the method according to the invention ;

- Figure 2 a diagram of the carrier discriminator; and

- Figure 3, a diagram of the code discriminator.

In the diagram of the decoder shown in FIG. 1, the code 101 and carrier 102 oscillators are controlled by the error signals which are described below. They make it possible respectively to obtain a point code and a Δ code in a generator 103 and sin signals. and cos. in phase shifters 104 and 105.

The punctual signals, Δ, cos. and sin. are applied to a correlator 106 which makes it possible, from the received signal, to obtain the signals l _p , l _Δ , Q _p and Q _Δ in a conventional manner using a set of multipliers 107 and blocking samplers 108.

For further processing these signals are sampled at a given frequency. According to the invention, in order to be able simultaneously to maintain the performance of the demodulation loops both in dynamic and in robustness and to correctly decode the data signals by letting the viterbi algorithm operate, two demodulators are used, a first demodulator of the DBPSK type for the carrier and a second demodulator of the DBPSK type for the data.

In the first loop, the signals sampled at the output of the correlator 106 are multiplied together in a module 109 to obtain crossed products l _p . I _Δ , l QΔ _> Qp • I _Δ and Q _p . Q _Δ . These cross products are integrated into a set of blocking integrator circuits 110, to then obtain the variables A, B, C and D defined by the formulas:

A = ∑I _P .I _Δ (1)

B = ∑I _P .Q _Δ (2)

C = ∑Q _P .I _Δ (3)

D = ΣQ _p .Q _Δ (4)

According to the invention, the signals l _p and Q _p are subjected to a so-called "squaring" treatment, which consists in performing a complex product l _p x / Q _p . This eliminates the data, and therefore their influence, and however introduces quadratic losses and doubles the dynamics of the signal. This operation removes the limitation of the predetection band coming from the data bits. It is performed in a complex square module 1 11 which therefore delivers the two signals l ' _p = l _p ² - Q _p ² and Q' _p = 2 l _p x Q _p . These signals are themselves integrated into a second set of blocking integrators 112.

The six signals thus obtained by the previous operations are sampled at a loop frequency, then they are combined together in a discriminator of code 113 to obtain a signal ε determined by the formula: ε = (cos ² + sin ² φ.D) l Jr ² _p + Q ' ² _p (5) this operation gives a gain of 3 dB on the precision of the code. The signal ε is then applied to a code corrector 114 which performs a conventional filtering operation. After sampling at the loop frequency, this filtered signal ε forms the error signal applied to the input of the code oscillator 101. Furthermore, the signals l _p 'and Q _p ' coming from "squaring" are applied to the discriminator carrier type DBPSK 115 which will be described later.

The signal φ coming from this discriminator 115 is applied on the one hand to the code discriminator 113, and on the other hand to a carrier corrector 116, which also performs a simple known filtering operation.

The signal φ is thus filtered and then sampled at the loop frequency to form the error signal of the carrier oscillator 102.

In an exemplary embodiment of the carrier discriminator of the DBPSK type, the diagram of which is shown in FIG. 2, the signals l _p 'and Q _p ' are first applied to a differentiator 201.

In this differentiator, an element 211 transforms it into Z ^"

¹ of these signals l _p 'and Q _p '. The two signals from this transform are applied, after multiplication by -1 for one of them in a multiplier 221, to a complex product computer 231, which also directly receives the signals l _p 'and Q _p '.

The signals at the output of this differentiator are then applied to a BPSK 202 type discriminator which performs the tangent arc operation (x, y) in an interval - π + π.

The output of this discriminator 202 is then applied to an integrator 203, which operates conventionally with an adder 213 and a generator 223 with a function in Z ^"1 .

We thus measure Δ φ _n , which is between - π and + π, and Δ φ _n + ι is then obtained by the formulas:

Aφ _{n + 1} = argument (l _{n + 1} + iQ _{n + 1} ). - argument ^ + iQ _n )

= argument (l _{n + 1} + iQ _{n + 1} ) ./ (l _n + iQ _n ) (6)

= argument (l _{n + 1} + iQ _{n + 1} ). (l _n -iQ _n ) According to the invention, the identification of the data bits is done in an identifier 117 from the signals l _p and Q _p supplied by the correlator 106. This identifier operates on the basis of a BPSK type discriminator. The phase discrimination is therefore then limited between - π and + π and the time constant Te / K is chosen small enough to decorrelate the successive bit errors.

The disadvantage of this method resides in a lower resistance to dynamics, which is largely compensated by the fact that the algorithm of Viterbi can function correctly.

In an exemplary embodiment, represented in FIG. 3, the signals l _p and Q _p are applied to a complex multiplier 301, which also receives the signals at the output of a module 302 which will be described later. The signals I and Q coming from this complex conduit are then applied to a discriminator 303 of the BPSK type, which performs the tangent arc operation (x, y) on a signal bounded between - π / 2 and + π 12.

The signal at the output of this discriminator is multiplied by a constant K in an amplifier 304, then it is applied to an integrator 305 in the structure is identical to that of the integrator 203 of FIG. 2.

The signal φ at the output of this integrator is applied to the module 302 which performs the operation e ^"1φ to give the signals provided above.

The data bits are then obtained by determining, using a comparator circuit 306, the sign of the signal I leaving the complex product generator 301.

Claims

1 - Method for processing a navigation signal containing data, in which a first carrier loop (101, 102,103,104,105,106,113,114,115,116) of the DBPSK type is used to demodulate the navigation signals, characterized in that a second BPSK loop (117) to identify the data bits.

2 - Method according to claim 1, characterized in that one uses in the first carrier loop a technique of "squaring" (111).

3 - Method according to claim 2, characterized in that one carries out a correlation (106) on the received signal to obtain signals l _p , l _Δ , Q _p and Q _Δ , which is carried out on these signals cross products (109) defined by the formulas:

Λ = ∑r _p J _&

C ^≈ ∑Q _p J _to

D = ∑Q _p .Q _to

that the "squaring" technique makes it possible to obtain signals l _p ² - Q _p ² and 2 l _p . Q _p , and that a code discrimination (113) is carried out from the crossed products and signals obtained by "squaring" by performing the operation given by the formula: ε = (cos ² + sinç? .cosf? + sin ² φ.D) l Jl ' _p + Q' _p