CN104062667A

CN104062667A - GPS weak signal tracking system based on I/Q branch correlation integral observation filtering

Info

Publication number: CN104062667A
Application number: CN201410314405.2A
Authority: CN
Inventors: 沈锋; 李伟东; 马娜娜; 韩浩; 李强; 桑静; 迟晓彤; 张金丽; 周阳; 兰晓明
Original assignee: Harbin Engineering University
Current assignee: Harbin Engineering University
Priority date: 2014-07-03
Filing date: 2014-07-03
Publication date: 2014-09-24

Abstract

The invention discloses a GPS weak signal tracking system based on I/Q branch correlation integral observation filtering. A receiver is used for receiving a satellite signal which is converted into an intermediate frequency signal through down conversion and transmitted to a frequency mixer. The frequency mixer is also used for receiving local sine and cosine reproduction carrier signals outputted by a local carrier digital-controlled oscillator, and outputting a I-branch output signal and a II-branch output signal to a correlation arithmetic unit. The correlation arithmetic unit also receives a local advanced C/A code, an instant reproduction C/A code and a lagged reproduction C/A code, wherein the local advanced C/A code, the instant reproduction C/A code and the lagged reproduction C/A code are generated by a code generator. Results are outputted to an integral eraser, and six-way relevant integral values are generated and outputted to a Kalman filter. The Kalman filter obtains an estimated carrier phase difference, an estimated carrier frequency difference and an estimated code phase difference, and the estimated carrier phase difference, the estimated carrier frequency difference and the estimated code phase difference are transmitted to the local carrier digital-controlled oscillator and the code generator. The system effectively reduces the noise intensity of the satellite signal, can trace the satellite signal better in a weak-signal environment, and improves the tracing precision.

Description

GPS weak signal tracking system based on I/Q branch correlation integral observation filtering

Technical Field

The invention belongs to the field of GPS signal tracking, and particularly relates to a GPS weak signal tracking system based on I/Q branch correlation integral observation filtering.

Background

The signals received by the GPS receiver from the satellite are spread spectrum modulation signals, and the navigation message can be obtained only by despreading and demodulating the satellite signals in the acquisition and tracking stages. In the GPS signal tracking stage, a signal channel starts from the rough estimated values of the carrier frequency and the code phase of the current satellite signal obtained in the acquisition stage, and the two signal parameters are gradually and finely estimated through a tracking loop. Under a weak condition, the tracking requirement cannot be met by simply increasing the integration time, and the situation that the loop integration time crosses the data bit edge of the navigation message can be avoided by using a bit synchronization method. Multipath effect can be eliminated by using Extended Kalman Filtering (EKF) during signal tracking, and the model is simple in structure but cannot effectively track weak signals. By adopting a fitting means at the triangular wave peak point of the signal autocorrelation function characteristic, the discontinuity of the Jacobian equation in the EKF operation process can be avoided, but the tracking error is larger.

Disclosure of Invention

The invention aims to provide a GPS weak signal tracking system based on I/Q branch correlation integral observation filtering, which can effectively track GPS weak signals.

The invention is realized by the following technical scheme:

a GPS weak signal tracking system based on I/Q branch correlation integral observation filtering comprises a receiver, a local carrier numerical control oscillator, a frequency mixer, a code generator, an integral cleaner, a correlation arithmetic unit and a Kalman filter,

the receiver is used for receiving satellite signals, converting the satellite signals into intermediate frequency signals through down-conversion and transmitting the intermediate frequency signals to the mixer;

the mixer also receives local sine and cosine recurrence carrier signals output by the local carrier numerically-controlled oscillator, performs frequency mixing operation on the intermediate frequency signal and the local sine recurrence carrier signal to obtain an I branch output signal, transmits the I branch output signal to the correlation arithmetic unit, performs frequency mixing operation on the intermediate frequency signal and the local cosine recurrence carrier signal to obtain a Q branch output signal, and transmits the Q branch output signal to the correlation arithmetic unit;

the correlation arithmetic unit also receives the local advanced reproduction C/A code, the instantaneous reproduction C/A code and the delayed reproduction C/A code generated by the code generator, and carries out correlation operation on the I branch output signal and the Q branch output signal respectively with the local advanced reproduction C/A code, the instantaneous reproduction C/A code and the delayed reproduction C/A code, and the result is transmitted to the integral clearing unit;

the integral cleaner generates six-path correlation integral value I according to the received information_E，I_P，I_L，Q_E，Q_P，Q_LAnd then transmitted to a Kalman filter; the Kalman filter obtains estimated carrier phase difference, carrier frequency difference and code phase difference, transmits the carrier phase difference and the carrier frequency difference to a local carrier numerical control oscillator, and transmits the code phase difference to a code generator;

the local carrier wave numerically controlled oscillator generates local sine and cosine reproduction carrier wave signals according to the received information and transmits the signals to the frequency mixer; a code generator generates a local advanced reproduction C/A code, an instantaneous reproduction C/A code and a delayed reproduction C/A code based on the received information, and transmits them to a correlation operator.

The GPS weak signal tracking system based on I/Q branch correlation integral observation filtering of the invention can also comprise:

1. the intermediate frequency signal of the ith satellite signal output by the receiver is:

wherein A is the normalized signal amplitude, D (k) is the data code, C (k) is the C/A code played by the satellite, f₁For the input frequency, f, of the radio-frequency signal_dFor signal doppler shift, k is epoch time,is the first carrier initial phase;

local sine reproduction carrier signal S generated by local carrier digital controlled oscillator_OSAnd a local cosine reproduction carrier signal S_OCComprises the following steps:

wherein A is_OIs the amplitude, f_OIn order to be the frequency of the radio,in order to be the phase position,

intermediate frequency signal and local sinusoidal reproduction carrier signal S_OSThe result of doing the mixing budget is:

wherein,

f_e＝f₁+f_d-f_O

intermediate frequency signal and local cosine reproduction carrier signal S_OCThe result of doing the mixing budget is:

2. the correlation operation method in the correlation operator comprises the following steps:

wherein C (k) is C/A code in output signal of the mixer, C_O(k) For a locally advanced or instantaneous or late reproduction C/a code generated by the code generator, N is the number of discrete data points participating in the correlation operation.

3. The state quantities of the kalman filter are:

wherein A is the normalized signal amplitude,is the carrier phase difference, δ w is the carrier frequency difference, δ a is the carrier frequency rate of change, δ τ is the code phase difference, λ_LFor the satellite signal carrier wavelength, λ_CAIs C/A code wavelength, W_kIs process noise, also known as system noise, noted

W_kIs a zero mean white noise sequence;

output six-way correlation integral value I of integral remover_E，I_P，I_L，Q_E，Q_P，Q_LFor the observations of the kalman filter:

where δ is the lead-lag interval of the locally reproduced C/A code, R (ε)_i) For locally reproducing the autocorrelation function of the C/A code, V_kTo measure noise.

4. The UT transform is adopted in the Kalman filter to generate a Sigma point chi and a weighting sequence w of symmetrical sampling:

<math> <mrow> <msub> <mi>χ</mi> <mn>0</mn> </msub> <mo>=</mo> <mover> <mi>x</mi> <mo>&OverBar;</mo> </mover> </mrow> </math>

<math> <mrow> <msub> <mi>χ</mi> <mi>i</mi> </msub> <mo>=</mo> <mover> <mi>x</mi> <mo>&OverBar;</mo> </mover> <mo>+</mo> <msubsup> <mrow> <mo>(</mo> <msqrt> <mrow> <mo>(</mo> <mi>n</mi> <mo>+</mo> <mi>λ</mi> <mo>)</mo> </mrow> <mi>P</mi> </msqrt> <mo>)</mo> </mrow> <mi>i</mi> <mi>T</mi> </msubsup> <mo>,</mo> <mi>i</mi> <mo>=</mo> <mn>1,2</mn> <mo>,</mo> <mo>·</mo> <mo>·</mo> <mo>·</mo> <mo>,</mo> <mi>n</mi> </mrow> </math>

<math> <mrow> <msub> <mi>χ</mi> <mrow> <mi>i</mi> <mo>+</mo> <mi>n</mi> </mrow> </msub> <mo>=</mo> <mover> <mi>x</mi> <mo>&OverBar;</mo> </mover> <mo>-</mo> <msubsup> <mrow> <mo>(</mo> <msqrt> <mrow> <mo>(</mo> <mi>n</mi> <mo>+</mo> <mi>λ</mi> <mo>)</mo> </mrow> <mi>P</mi> </msqrt> <mo>)</mo> </mrow> <mi>i</mi> <mi>T</mi> </msubsup> <mo>,</mo> <mi>i</mi> <mo>=</mo> <mn>1,2</mn> <mo>,</mo> <mo>·</mo> <mo>·</mo> <mo>·</mo> <mo>,</mo> <mi>n</mi> </mrow> </math>

wherein the state variable X_KMean value of initial distribution ofMean square error matrix is P, lambda ═ alpha²(n + κ) -n is a scalar quantity, α is used to control the distance of each point from the mean, 10^-4Alpha is more than or equal to 1, kappa is a scalar quantity, and scalar quantity beta is used for reducing the influence of high-order moment;

the time updating process of the Kalman filter is as follows:

χ_i,k|k-1＝f(χ_i,k-1|k-1)i＝1,2,…,2n

wherein, χ_i,k-1|k-1Is an estimate of the ith Sigma sample point at time k-1, f (-) is a non-linear function,predicting the state quantity prediction value at time k for time k-1, P_k|k-1Mean square error matrix, Q, for predicting state quantities at time k-1_kIs a covariance matrix of process noise;

the measurement updating process of the Kalman filter comprises the following steps:

P_ZZmeasuring mean square error matrix, P, for the values after updating_XZCross-correlation mean square error matrix, R, for state and quantity measurements after updating of measurement values_kTo measure a noise covariance matrix.

The invention has the beneficial effects that:

the invention can overcome the error caused by the output of the traditional discriminator, can accurately model the nonlinear system, effectively reduce the noise intensity of satellite signals, can better track the satellite signals in the weak signal environment and improve the tracking precision. The I/Q branch correlation integral value of the receiver is used as the most original data of the tracking loop, so that the error caused by the output of the traditional discriminator is overcome; the influence of thermal noise and dynamic stress error in the receiver is reduced, and the phenomenon of tracking loop lock losing is reduced. And a UKF filtering algorithm is applied to process and filter the loop, so that the nonlinear relation between the output of the traditional receiver discriminator and the state quantity is overcome. The processing of the nonlinear signal is fast in convergence, and the dynamic parameters can be accurately estimated.

Drawings

FIG. 1 is a diagram of a tracking loop based on I/Q branch correlation integral observation filtering;

FIG. 2 is a block diagram of a Kalman filter tracking loop.

Detailed Description

The present invention will be described in further detail with reference to the accompanying drawings.

The core technology of GPS signal tracking based on I/Q branch correlation integral observation filtering is to use I/Q branch correlation integral to construct a centralized filtering model and apply UKF filtering algorithm, the basic idea is to use the I/Q branch correlation integral of a receiver tracking loop as observed quantity, use a centralized filter to replace a discriminator in the traditional GPS tracking loop, and use UKF filtering algorithm to jointly estimate the characteristic quantity in a carrier loop and a code loop, so as to achieve the accurate tracking of parameters such as carrier frequency, carrier phase, code phase and the like. Because the correlation integral value of the I/Q branch of the receiver and the state quantity in the centralized filter have a nonlinear relation, the problem can be well overcome by adopting a UKF filtering algorithm.

The system modeling accuracy is a key technical problem faced by GPS signal tracking based on the I/Q branch correlation integral value as the observed quantity, and the selection of the state quantity is particularly important for the system modeling accuracy because the system observed quantity and the state quantity to be estimated have a nonlinear relation. For a nonlinear system, the increase of the dimension of the state quantity can more accurately reflect the information of each state of the system, and meanwhile, the calculation quantity is correspondingly increased. In order to simplify the system structure and reduce the system computation, parameters such as carrier phase, carrier frequency and code phase are usually selected as the state quantities of the system when establishing the model.

As shown in fig. 1, a satellite signal received by a receiver through an antenna is down-converted into an intermediate frequency signal, the intermediate frequency signal is subjected to frequency mixing operation with a local carrier signal generated by a local carrier Numerically Controlled Oscillator (NCO), the frequency mixing result is subjected to six-path correlation operation with a locally reproduced C/a code generated by a code generator, and the six-path correlation operation is used as an input signal of a centralized filter to participate in UKF filtering after passing through an integrating-removing device, and the filtering result is fed back to the local NCO, so that the loop is closed.

Fig. 2 is a diagram of a tracking loop structure of a centralized filter, and it can be seen from the diagram that after the intermediate frequency signals of n tracking channels are subjected to carrier stripping and code correlation operation, the intermediate frequency signals are respectively subjected to an integral-cleaner to eliminate high frequency signal components and noise in I/Q branch signals so as to improve a carrier-to-noise ratio, and then enter the centralized filter, and the I/Q branch correlation integral of each channel is used as an observed quantity of the UKF filtering.

The invention specifically comprises the following steps:

step one, carrier stripping and code correlation operation;

the receiver utilizes the satellite signal received by the antenna to send to the radio frequency front-end processing, converts the radio frequency signal into an intermediate frequency signal through multi-stage frequency mixing, and carries out frequency mixing operation with the local sine and cosine reproduction carrier signal generated by the local carrier digital controlled oscillator, so that the intermediate frequency carrier including Doppler frequency shift in the input signal is thoroughly stripped, the processed intermediate frequency signal carries out correlation operation with the advanced, instantaneous and delayed six paths of C/A codes generated by the C/A code generator, and six paths of correlation integral values are formed after passing through the integral-cleaner.

The ith satellite signal output by the receiver rf front end can be written as:

a is normalized signal amplitude, D (k) is data code, C (k) is C/A code played by satellite, f₁For the input frequency, f, of the radio-frequency signal_dIn order to shift the doppler of the signal,k is the time of the epoch, k is the epoch time,the first carrier initial phase.

The local sine and cosine reproduction signals generated by the local carrier NCO are:

in the formula, A_OFor local reproduction of the signal amplitude, f_OIn order to locally reproduce the frequency of the signal,the signal phase is reproduced locally.

The intermediate frequency signal S and the local sinusoidal reproduction signal S are combined_OSThe mixed loop is branched into I branch to obtain intermediate frequency signal S and local cosine reproduction signal S_OCThe loop branch of the frequency mixing becomes the Q branch, where the intermediate frequency signal S is reproduced with the local sine on the I branch_OSWhen multiplying and mixing operation is carried out in a mixer, the product i is obtained_p(k) Is composed of

Wherein,

f_e＝f₁+f_d-f_O （4）

in the formula (3), the first term on the right of the equal sign is a low-frequency component, the second term is a high-frequency component, and f_eAndrespectively, an intermediate frequency signal S and a local sinusoidal reproduction signal S_OSThe carrier frequency difference and the carrier phase difference between them.

Mixing result i_p(k) Filtering out high frequency component by low pass filter to obtain the following filtering result

And the state quantity error information fed back by the Kalman filter is sent to the carrier NCO to update and output a local reproduction signal, so that the phase time between the local reproduction signal and the input intermediate frequency signal is kept consistent. F in equation (5) when the phases of the input and output signals are substantially coincident_eAndtowards zero value, i_p(k) Only the signal amplitude, the data code D (k) and the pseudo code C (k) are left, so that the carrier stripping effect is completely achieved.

Similarly, the intermediate frequency signal S can be derived on the Q branch and the local cosine reproduction signal S_OCThe mixing result is obtained by filtering the high frequency component by a low pass filter

The C/A code in the received signal is C (k), and the locally reproduced C/A code C_O(k) Performing a correlation operation to obtain the following correlation results:

where N is the number of discrete data points participating in the correlation operation, typically corresponding to a 1ms long acquisition data volume.

Mixing result I on branch I_p(k) Respectively and simultaneously correlating with the local early, real-time and late reproduction C/A codes to generate i_E，i_PAnd i_LThe epoch time k in the signal is omitted for simplicity of representation. Mixing result Q on branch Q_p(k) Also respectively and simultaneously performing correlation operation with the three C/A code signals to generate q_E，q_PAnd q is_LA signal. At this time, C/A code in the intermediate frequency signal is thoroughly stripped, and the despread i_E，i_P，i_L，q_E，q_PAnd q is_LOnly contains the data code information D (k) which is wanted to be obtained, and the six paths of related integral values I are formed after the data code information D (k) passes through the integral-elimination device_E，I_P，I_L，Q_E，Q_PAnd Q_LAs an observed quantity of the kalman filter, as shown in fig. 1.

Secondly, observing a filtered signal tracking model based on coherent integration;

selecting normalized signal amplitude A and carrier phase differenceAnd the carrier frequency difference delta w, the carrier frequency change rate delta a and the code phase difference state delta tau are state quantity column writing state equations.

And taking the six paths of correlation integral values of the I/Q branch, which are advanced, real-time and delayed, as observed quantities of UKF filtering, and writing a measurement equation.

The GPS tracking loop is essentially a digital phase-locked loop, the angular analysis from a control system is a classical phase tracking control system, the Kalman filtering algorithm is an optimal estimation method using a minimum variance criterion, the state parameters of the system can be accurately estimated, the system can be more effectively controlled, and in order to accurately track the frequency and the phase of an input signal, the state quantity of a filter at the moment k is selected as follows:

wherein A is the normalized signal amplitude,for the carrier phase difference, δ w is the carrier frequency difference, δ a is the carrier frequency change rate, and δ τ is the code phase difference. The state equation of the kalman filter is as follows:

in the formula, λ_LIs a satellite signal L₁Carrier wavelength, λ_CAIs C/A code wavelength and has a value of about 293m, W_kIs process noise, also called system noise, noted as:

W_kis a zero mean white noise sequence.

And (3) taking the I/Q branch six-path correlation integral output value in the step one as the observed quantity of the Kalman filter, and establishing a nonlinear equation related to the state quantity as follows:

where δ is the lag-lead interval of the local C/A code, R (ε)_i) Is the autocorrelation function of the C/A code. V_kTo measure noise, the variance matrix of the noise is:

in the formula, σ_NIs the noise strength after processing of the correlation signal I, Q.

Due to the fact that the observation equation and the state quantity have a nonlinear relation, the application of the traditional Kalman filtering algorithm can cause poor estimation precision, and therefore the UKF filtering algorithm is adopted to carry out loop processing filtering, and the estimated carrier phase difference is obtained finallyThe carrier frequency difference δ w and the code phase difference δ τ.

Step three, Kalman filter based on UKF

And a UKF filtering algorithm is adopted to replace a loop filter, so that the contradiction between the dynamic property and the filtering precision of the traditional tracking method is eliminated. The strong nonlinear tracking capability of UKF is utilized to replace a discriminator to accurately track the phase, frequency and code phase of multi-channel carrier waves in a weak signal environment.

The UKF filtering algorithm adopts effective combination of UT transformation and traditional Kalman filtering framework, and obtains state estimation through estimation of nonlinear function probability density function, and the UT transformation is the core and the basis of the UKF algorithm. The idea of the UT transform is: ensuring that the mean and covariance of the samples areAnd on the premise of P, selecting 2n +1 point sets (Sigma point sets), wherein n is a state variable dimension, applying nonlinear transformation to each sampled Sigma point to obtain point sets subjected to nonlinear transformationAnd p_yIs the statistic of the transformed Sigma point set.

The nonlinear system model after the discretization of the Kalman filter is as follows:

\begin{matrix} X_{K + 1} = f (X_{K}) + W_{K} \\ Z_{K + 1} = h (X_{K}) + V_{K} \end{matrix} - - - (12)

state variable X_KMean value of initial distribution ofThe mean square error matrix is P. The UT transform produces a Sigma point χ and a symmetrically sampled weighted sequence w, expressed as:

<math> <mrow> <msub> <mi>χ</mi> <mn>0</mn> </msub> <mo>=</mo> <mover> <mi>x</mi> <mo>&OverBar;</mo> </mover> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>13</mn> <mo>)</mo> </mrow> </mrow> </math>

<math> <mrow> <msub> <mi>χ</mi> <mi>i</mi> </msub> <mo>=</mo> <mover> <mi>x</mi> <mo>&OverBar;</mo> </mover> <mo>+</mo> <msubsup> <mrow> <mo>(</mo> <msqrt> <mrow> <mo>(</mo> <mi>n</mi> <mo>+</mo> <mi>λ</mi> <mo>)</mo> </mrow> <mi>P</mi> </msqrt> <mo>)</mo> </mrow> <mi>i</mi> <mi>T</mi> </msubsup> <mo>,</mo> <mi>i</mi> <mo>=</mo> <mn>1,2</mn> <mo>,</mo> <mo>·</mo> <mo>·</mo> <mo>·</mo> <mo>,</mo> <mi>n</mi> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>14</mn> <mo>)</mo> </mrow> </mrow> </math>

<math> <mrow> <msub> <mi>χ</mi> <mrow> <mi>i</mi> <mo>+</mo> <mi>n</mi> </mrow> </msub> <mo>=</mo> <mover> <mi>x</mi> <mo>&OverBar;</mo> </mover> <mo>-</mo> <msubsup> <mrow> <mo>(</mo> <msqrt> <mrow> <mo>(</mo> <mi>n</mi> <mo>+</mo> <mi>λ</mi> <mo>)</mo> </mrow> <mi>P</mi> </msqrt> <mo>)</mo> </mrow> <mi>i</mi> <mi>T</mi> </msubsup> <mo>,</mo> <mi>i</mi> <mo>=</mo> <mn>1,2</mn> <mo>,</mo> <mo>·</mo> <mo>·</mo> <mo>·</mo> <mo>,</mo> <mi>n</mi> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>15</mn> <mo>)</mo> </mrow> </mrow> </math>

in the formula,is the ith row (or column) of the root mean square matrix obtained by Cholesky decomposition; λ ═ α²(n + κ) -n is a scalar quantity, α is used to control the distance of each point from the mean, and is typically in the range of (10)^-4Alpha is more than or equal to 1); κ is also a scalar, usually taken to be 0; the function of the scalar β is to reduce the effect of high order moments, and when the state quantities are gaussian distributed, the optimal value is 2.

And (3) time updating process:

χ_i,k|k-1＝f(χ_i,k-1|k-1)i＝1,2,…,2n (18)

where the estimate χ at time k-1 of the ith Sigma sample_i,k-1|k-1Obtaining k-1 time prediction k time estimation value χ after nonlinear function f (-) conversion_i,k|k-1，χ_i,k|k-1And weight of the ith Sigma sample PointObtaining the state quantity predicted value of k-1 moment predicted k moment after operation

In the formula, P_k|k-1Mean square error matrix, Q, for predicting state quantities at time k-1_kIs a covariance matrix of process noise.

And (3) measurement updating process:

in the formula, P_ZZMeasuring mean square error matrix, P, for the values after updating_XZCross-correlation mean square error of state quantity and quantity measurement after updating of measurement valueDifferential array, R_kTo measure a noise covariance matrix.

K_k＝P_XZ(P_ZZ)^-1 (23)

{\hat{X}}_{k | k} = {\hat{X}}_{k | k - 1} + K_{k} (Z_{k} - {\hat{Z}}_{k | k - 1}) - - - (24)

P_k|k＝P_k|k-1-K_kP_ZZK_k ^T (25)

In the formula,optimum estimated value of state variable at time K after updating of measured value, K_kFor filtering the gain matrix, P_k|kIs a mean square error matrix of the state quantity at the k moment.

The Kalman filter adopts a UKF filtering algorithm, selects a series of Sigma sampling points in the time updating process, and synthesizes the sampling points after nonlinear transformation according to respective weights to obtain an estimated value of the system state quantity at the current moment after passing through a state equation and a measurement equation. In the measurement updating process, the estimated value of the state quantity of the Kalman filter is corrected by utilizing the six paths of related integral values, the corrected carrier frequency, carrier phase and code phase difference are sent to a local carrier and code NCO, and locally reproduced carrier and code output signals are updated in real time, so that a complete closed loop is formed.

Claims

1. GPS weak signal tracking system based on I/Q branch correlation integral observation filtering, its characterized in that: the satellite signal acquisition device comprises a receiver, a local carrier numerical control oscillator, a mixer, a code generator, an integral cleaner, a correlation arithmetic unit and a Kalman filter, wherein the receiver is used for receiving a satellite signal, converting the satellite signal into an intermediate frequency signal through down-conversion and transmitting the intermediate frequency signal to the mixer;

the local carrier wave numerically controlled oscillator generates local sine and cosine reproduction carrier wave signals according to the received information and transmits the signals to the frequency mixer;

a code generator generates a local advanced reproduction C/A code, an instantaneous reproduction C/A code and a delayed reproduction C/A code based on the received information, and transmits them to a correlation operator.

2. The GPS weak signal tracking system based on I/Q branch correlation integral observation filtering according to claim 1, characterized in that: the intermediate frequency signal of the ith satellite signal output by the receiver is:

wherein,

f_e＝f₁+f_d-f_O

intermediate frequency signal and local cosine reproduction carrier signal S_OCAs a result of the mixing budget：

3. The I/Q branch correlation integral observation filtering based GPS weak signal tracking system according to claim 2, wherein: the correlation operation method in the correlation operator comprises the following steps:

4. The GPS weak signal tracking system based on I/Q branch correlation integral observation filtering according to claim 3, characterized in that: the state quantities of the kalman filter are:

W_kIs a zero mean white noise sequence;

5. The GPS weak signal tracking system based on I/Q branch correlation integral observation filtering according to claim 4, characterized in that: the UT transform is adopted in the Kalman filter to generate a Sigma point chi and a weighting sequence w of symmetrical sampling:

the time updating process of the Kalman filter is as follows:

χ_i,k|k-1＝f(χ_i,k-1|k-1)i＝1,2,…,2n

wherein, χ_i,k-1|k-1Is an estimate of the ith Sigma sample point at time k-1F (-) is a non-linear function,prediction of state quantity at time k, P_k|k-1Mean square error matrix, Q, for predicting state quantities at time k-1_kIs a covariance matrix of process noise;