DK166468B1 - Digital receiver for satellite navigation system with Doppler adjustment - Google Patents

Description

in DK 166468 B1

The present invention relates to a receiver for the Navstar satellite in the navigation system.

Navstar is a satellite navigation system that is designed to provide continuous worldwide weather coverage, providing very accurate three-dimensional position and velocity information.

The complete Navstar system is planned to consist of 18 satellites located in nearly circular orbits with radii of 26,600 km and an inclination to the Earth's equatorial pi an of 55 degrees.

Each satellite emits two navigation signals denoted LI and L2 and centered on 1575 and 1228 MHz respectively.

Both signals transmit distance information using time-locked modulations to atomic standards. The forms of these modulations (known as pseudo-random codes because they occur randomly but are nonetheless well-defined) are peculiar to each satellite.

By measuring the phases of the received codes relative to a clock transmitter in the receiver along with the doppler displacements of the radio frequency carriers, a user can calculate the distance and range rate of a particular satellite by monitoring four satellites (Fig. 1 ). By decoding data about their movements, which are also modulated on the transmitted signals, the user can solve equations (Fig. 2) to determine his three-dimensional position and velocity and also make corrections to his 25 clock sensor to make it in accordance with satellite time.

Alternatively, if the user is restricted to moving on the Earth's surface or at a known altitude, he can perform two-dimensional measurements using three satellites.

30 The software that controls the receiver must select from among the satellites within sight the subset that provides the most favorable geometry for the navigation calculations.

In fact, two pseudo-random codes are broadcast by each satellite. The first of these is used to assist in the collection of the satellite signals and to provide rough navigation and is thus called the Coarse / Acquisition code C / A. The other has a ten times higher modulation speed, which gives the full navigation accuracy of the system and is referred to as the precision code P code.

2 DK 166468 B1

A basic Navster receiver typically includes a low noise amplifier and down-converter to an appropriate intermediate frequency followed by one or more code and carrier channel channels, each capable of tracking the transmissions from one or the other satellite. There are also 5 associated distance and distance change speed measurement circuits.

The purpose of the code tracking loop is to keep a code generator in the receiver in step with a received pseudo-random sequence (sequence), thus providing information about the distance to the satellite being followed.

10 In order to obtain a position and speed estimate, the receiver must be locked to the transmissions from a number of satellites. The case of a complete three-dimensional estimate for which the required number is four is considered, as shown in FIG. 1. Four "pseudo-distance" measurements are performed by locking code tracking loops to the 15 received signals and then timing the occurrence of certain states of the code generators within the loops by the receiver's clock sensor. The measurements are of "pseudo-distance" rather than true distance due to the (as yet) undetermined receiver rate transducer displacement.

Similarly, by measuring the frequencies of the carrier-tracking loop voltage-controlled oscillators over keying times determined by the receiver's clock sensor, four measurements of "pseudo-distance-change rate" are obtained. These are flawed compared to the true distance change rates due to the frequency sensor's frequency error. All of these measurements, together with data from each satellite, which provides satellite motion information, then enable a navigation solution to be obtained. This is based on the fact that four observations are required to solve for four unknowns.

According to the present invention, there is provided a receiver for a Navstar satellite navigation system including means for amplifying and down-translating to mf frequencies to generate quadrature signals, analog-to-digital transducers to separately digitize the quadrature signals, local digital code generator organs, means for of the digitized quadrature signals separately with the same locally generated digital codes, channel signal processing means to which the output signals from the correlation means are provided, which processing means are arranged to control the code and carrier sequence of the receiver, and correction means are provided depending on the signal. produced in the processing means to cause phase rotation of the baseband signal phasor represented by the quadrature signals to cause doppler tracking in the receiver loop, and it is peculiar to the receiver 5 of the invention that the receiver includes a digital data storage means charged with data in the form of "lookup" tables including combinations of signal input states, means for supplying the digitized quadrature signals, as partial address signals for the storage means, and means 10 for generating additional address signals for the storage means in dependence on control signals, whereby phase rotation correction is performed by accessing the storage means in accordance with the address signals to produce an output signal for channel signal processing in the input means.

FIG. 3 shows a generalized Navstar receiver architecture. Signals are introduced on L-bands and are passed through successive steps for amplification and downconversion at rf, mf and zero mf frequencies.

At some point in the chain, the signal must pass through an analog-digital interface to allow information extraction using a digital processor. If the code and carrier loops are closed in software, this processor will also provide the necessary feedback control signals.

It has hitherto been known to perform the necessary Doppler correction of the digital signals in the rf portions of the receiver.

25 There are a number of possible places where code and carrier (Doppler) adjustment can take place: by mf baseband analogue or baseband digital. After the alignment point in the receiver chain, the circuit is dedicated to receiving signals from a particular satellite. Accordingly, to receive broadcasts from more than 30 satellites, the circuit after that location must be repeated corresponding to the intended number of satellites or alternatively time-shared (cyclic or multipiex) between the same number. Therefore, to reduce circuit complexity, the adjustment site should be pushed as far back into the chain as possible. The farthest place where this process can be performed is by performing both code correlation and Doppler correction on digital baseband.

Other considerations can also be made in favor of a basic tape solution. When performing code correlation on baseband, true multipliers can be used instead of mixers, thus avoiding the problem of mixer imperfections. The stability and Q-factor of the required filters to determine the pre-correlation bandwidth would require rather stringent specifications at mf. The problem is greatly alleviated by performing low pass filtering on baseband.

5 The need to use multiple transfer loops in the synthesizer to implement mf doppler insertion can also be avoided.

The invention will now be explained in more detail with reference to the drawing, in which fig. 1-3 show the prior art mentioned above; 4 shows the effect of Doppler displacement; FIG. 5, phase rotation FIG. 6 is a numerically controlled oscillator and FIG. 7 a channel in a Navstar receiver.

The possibility of providing digital Doppler correction on 15 basebands is highly desirable as this will allow the use of a single fixed frequency down-turnover to zero mf followed by a single pair of A / D converters regardless of the required number of receiver channels.

To represent the signal phasor on baseband, in-phase (I) and quadrature (Q) channels are required, where the I and Q channels represent the real and imaginary component of the phasor. Any doppler displacement will cause the phasor to rotate, thus producing a doppler loss if filtration is performed by accumulating successive phasor samples. This effect is shown in FIG. 4. The rotation must therefore be removed or significantly reduced, 25 before significant accumulation can take place.

The signal vector (phasor) can be expressed in exponential form such that S = AeJ (ωΝΤ + Φ) N = 0.1,2 ____ 30 where A is the signal amplitude, ω is the Doppler frequency, T the sample interval, and φ is any angle.

To remove the phase rotation, the signal vector must be multiplied by an opposite rotating unit vector thus 35 S · - AeJ <"NT + ♦>. E 'j" NT - Aej *

The phasor will now appear to be stationary and can accumulate in time without loss.

5 DK 166468 B1

The practical implementation of the opposite rotary function on the I and Q channels can be readily realized by expressing the multiplication in real and imaginary part as follows: 5 (I + jQ) (cos ω NT - j sin ω NT)

= I cos w NT + Q sin NT + jQ cos ω NT - jl sin ω NT

v --- J "" -v-r- 'Γ Q' 10 The transformation is shown schematically in FIG. 5th

Conveniently, the angle of rotation ω NT can be provided by a numerically controlled oscillator (NCO) of the form shown in FIG. 7. The angle will then be represented by an M bit digital word. The required word length will be determined by the maximum phase noise that can be tolerated from the rotation operation. The resulting phase noise will be given by evaluating the effective value of the quantization noise. If 6 bits are used, a phase quantization of 0.098 radians will be obtained with an associated effective value of phase noise σψ of 20 <7 2o = ° '098 12 which gives σ ^ - 0.028 rad. eff.

This value will typically be well below the expected 25 thermal noise in the Navstar phase tracking loops.

The frequency range and resolution of NCO should be sufficient to cover the complete expected doppler range in steps small enough to prevent significant phase errors from accumulating between NCO updates. A Doppler range of + 10 kHz will be more than 30, as it will enclose the full satellite Doppler range of ~ +4 kHz along with a user speed range of + Mach 3.8. Considering the frequency of the apparatus, it can be assumed that the NCO is updated at an effective rate of approx. twice the loop bandwidth. In the case of a narrow bandwidth of 35, a refresh rate of approx. Thus, 1 Hz, a frequency resolution of 0.01 Hz will allow a phase error in the worst case of - 0.05 radians to occur. This is compatible with the above phase noise. The number of bits required to control NCO is therefore determined as DK 166468 B1 6

Tog2 (20.103 / 0.01) ~ 21 bit.

The NCO must also be clock controlled at a sufficiently high speed to prevent "mouse teeth" on the phase ramp output from occurring. These "mouse teeth" are generated because the oscillator produces only a finite number of output samples per output cycle. Therefore, the problem becomes worst at the highest output frequency. To reduce this effect to the level of the phase quantization, approx. 64 output samples per output cycle therefore be -10 required. This corresponds to a clock speed of 640 kHz.

Positioning of the phasor rotator after any accumulation of the correlator output is acceptable, provided that there is no significant Doppler loss during this accumulation time. The loss can easily be evaluated by examining the frequency characteristic F (tt>) of the accumulator thus F (·) - I "eij-T.

N N sin ^ 20 1 = 1 2

Therefore, for a maximum 1 dB loss at the maximum Doppler frequency of 10 kHz, N must not be greater than 547. Placing the phasor rotator after this accumulation amount would result in the device's throughput rate being reduced from 20 MHz to approx. 40 kHz.

It is advantageous to perform some of the accumulation before the phase rotation, thereby reducing the ROM's access time requirements. Further accumulation can then be used to reduce the output data rate to a sufficiently low value for handling a microprocessor. This will be of the order of 1 kHz. However, there is a further aspect of the configuration that needs to be examined, namely the required I and Q word lengths.

The required number of bits for I and Q di g numbers in the series will be application dependent. If a 2 dB loss can be tolerated, single 35 bit conversion will be satisfactory. However, if 2 bits are used, this loss will be reduced to 0.6 dB. These two cases assume that the signal-to-noise ratio is negative. As you progress through the accumulation steps, this will not always be the case and more bits will be needed.

7 DK 166468 B1

The location at which the phase rotation is performed will therefore depend on the implementation of the device. A 2 bit rotator operating at 20 MHz can be placed directly before or after the granule. Alternatively, a slower but larger word length rotator can be used after a limited amount of post-correlation accumulation.

Considering the 2 bit 20 MHz resolution, it can be seen that a simple implementation could be achieved using a PROM lookup table. The addresses would consist of 4 bits of I and Q signal data along with the 6 bits needed to define the phase rotation to a resolution of 5.625 ° (360 °: 64, i.e. 6 bits in binary). The output signals would simply be 4 bits of rotated I and formationι nformation. Therefore, the PROM size would be 4 k done as a 1 k x 4. If single bit digitization was sufficient, this could be reduced to 256 x 2.

15 This idea can be taken a step further by simply combining the single multiplication required to implement the correlation function in the rotation PROM. Finally, the in-time correlation together with the switched or dedicated early / late correlation can also be combined. With regard to the switched early / late case, the 20 PROM would require 12 address lines and an output word size of 8 bits, thus establishing a 32 k PROM. This configuration is shown in FIG.

7th

The L-band input signal for all channels is first transmitted through a single chain of rf and mf steps to a frequency FQ + D.

The resulting mf signal is fed to two mixers 70a, 70b in quadrature with a local oscillator signal at a frequency FQ. The I and Q signals thus obtained are low pass filtered and fed to A / D converters 71a, 71b. Adaptive threshold methods are used to combat non-Gaussian noise sources. 2-bit digitization is used in conjunction with sampling early, late and prompt (e, l, p) correlations to address a programmable storage chip 72. The corresponding early, late or prompt output states are separately accumulated and fed to the channel processor 73. Measurements is performed by the channel processor of distance and distance change rate by monitoring the code generator modes. A numerically controlled oscillator 74 responds to channel processing and derives the required phase rotation angle as well as generates a 6 bit word which forms part of the PROM address along with the I and Q digitization and e, 1, p of the code sample.

8 DK 166468 B1

However, the primary function of the processor is to keep the code and carrier phases on track.

Estimates of code position errors can be made simply by forming the difference of early and late correlation samples. These 5 are derived by performing I + Q operations on the early and late output signals. It should be noted that in this digital implementation, channel balance will no longer be a problem.

Code position error estimates can then be passed to a software loop filter before being used to update the code generator, thus closing the code tracking loop.

Carrier phase estimator can be made using a Costas I.Q. technique on the prompt correlation samples. The carrier loop will then be closed in a similar manner to the code loop. Carrier frequency estimates can also be formed by performing an operation 15 on time sequential I, Q pairs as shown below Q.I. i - I.Q. .

Error frequency oc x 1 1 ~

Tf + Oi

This malfunction can be used to assist initial carrier phase attainment and can also be used to provide frequency estimates when severe interference precludes the use of carrier phase tracking loop.

This configuration allows the addition of multiple receiver channels in a simple manner by adding additional code generators NCOs and PROMS. The same A / D module and the same channel processor can be used for the extra channels. However, a separate A / D turnover module will be required if L1 and L2 are to be received simultaneously.

For a lower performance receiver channel, the adaptive threshold 30 2 bit A / D converters can be replaced with simple tbi units. The granular atoms need only be of a switched early / late type and thus require only a single pair of I and Q outputs.

If simultaneous operation on a number of satellites or on different signal segments of the same satellite is required, a number of 35 serial correlation blocks can be used in parallel.

Claims (5)

1. Receiver for a Navstar satellite navigation system including amplifying and down-switching means (70a, 70b) to 5 mf frequencies to generate quadrature signals, analog-to-digital transducers (71a, 71b) for separate digitization of quadrature signals, local digital code generation means, means for correlating the digitized square signals separately with the same locally generated digital codes, channel signal processing means (73), to which the output signals from the correlation means are fed, which processing means are adapted to control the code and carrier tracking of the receiver, and correction means which is adapted to cause phase rotation of the baseband signal phasor 15 represented by the quadrature signals in response to control signals produced in the processing means, to cause Doppler tracking in the receiver loop, characterized in that the receiver includes a digital data storage means (72), s "tables which include combinations of signal input states, means for supplying the digitized square signals as partial address signals for the storage means, and means (74) for generating additional address signals for the storage means in response to the control signals, whereby phase rotation correction is performed the storage means in accordance with the address signals to produce an output signal for the channel signal processing means. Receiver according to claim 1, characterized in that the receiver includes means for generating additional address signals for the storage means in dependence on the local digital code generating means, so that quadrature signal correlation 30 is also performed upon access to the storage means (72). Receiver according to claim 1 or 2, characterized in that the Doppler offset determined in the processing means (73) is a numerical value represented as a digital word, the correction means being a numerically controlled oscillator (NCO) (74) to which the digital word is applied. , the output signal of the NCO being said additional address signals of the storage means (72). Receiver according to claim 1,2 or 3, characterized in that the storage means is implemented as a programmable read action (PROM) (72). DK 166468 B1 10 Receiver according to any of the preceding claims, characterized in that the receiver includes means for generating additional address signals for the storage means (72) to effect correlation of the digitized quadrature signals 5 with the locally generated digital codes when they are locally generated. codes are "early" or "late" with respect to the time of receipt of the square turbulence. 10 15 20 25 30 35