JP3783022B2 - Receiver for wide-area position measurement system - Google Patents

Description

に乗じ、次の時間間隔に対する状態ベクトルの推定値を得る。このループは、状態遷移ループとして知られる。Ｎ回遷移するたびに、観察が行われ、

に対して補正が行われる。従って、観察値の間に、遷移マトリックスは受信機の予想できる動的挙動のモデルを作る。
【００４５】
予想された観察ベクトル

を観察ベクトルＹと比較し、その結果得られた予測エラー・ベクトルを処理することによって補正値が作られる。後者には、いわゆるカルマ−・ゲイン・マトリックスＫを乗じる。後者は、状態共分散マトリックスＰ、観察エラー共分散マトリックスＲ、システム雑音の共分散マトリックスＱ、及び観察マトリックスＭを含む確率的モデル技術によって計算する。この観察マトリックスは、測定ベクトルを状態ベクトルと関連づける。
【００４６】
モデル（例えば、φ、Ｑ、及びＭ）を形成するマトリックス上に記述子が存在するが、このことは、これらのマトリックスは時間の不変量であるので、定期的に更新しなければならないことを示している。これらのマトリックスは、衛星と受信機の位置によって決まる。更に、ＱとＲは、フィルタ６２が受信機の動的挙動の変化に適応するのを可能にする種々の方法によって観察値から推定することができる。フィルタ６２を適応できる他の方法は、当該予想エラーの推定値の軌跡に関するスカラー係数を乗じることによる方法である。詳しくは、スカラー係数は、１＋αであり、ここでαはマトリックス

のサブ軌跡と直線的に関連する。
【００４７】
カルマン・フィルタ６２に関連する観察は、疑似距離補間アルゴリズムからの入力と周波数と位相推定子からの入力によって行われる。所定数の観察が行われた後、位置の推定値は予想された状態ベクトルから、直接ユーザに出力できる。同様に、微分演算を行ってベース基地に送信しようとする疑似距離の差異の推定値は、予想された観察ベクトルから直接取り出すことができる。疑似距離補正値がベース基地から入手可能である場合、図３に示すように、フィルタ６２に供給することによってこれらの値を使用してより正確な位置を発生することができる。
【００４８】
位置測定のアルゴリズム
この受信機で使用する位置測定のアルゴリズムは、カルマン・フィルタ６２内で実行される。
【００４９】
本発明は、ここで説明した実施例に限定されないことを強調する。多くの変更と変形を、説明した広範な概念の範囲内で行うことができる。受信機の設計は幾つかの方法によって向上させることができる。
【００５０】
ここで説明した包括的な概念は、「多重チャンネル」受信機として知られる、４つ以上のチャンネルを有する受信機に関するものである。
【００５１】
殆ど努力を加えることなく、本概念は、３チャンネルの船舶用受信機または１チャンネル多重受信機のような用途に容易に適用することができる。１チャンネル受信機の場合、重要な相違点は、差異ではなく、疑似距離、位相等の絶対測定値を必要とされる点である。他の変更と変形を行って、特定の用途と特別の環境条件に対応することを意図する。
【図面の簡単な説明】
【図１】図１Ａ〜図１Ｄの配置を示す構成図である。
【図１Ａ】本発明による１つの受信機の機能ブロック図の一部である。
【図１Ｂ】本発明による１つの受信機の機能ブロック図の一部である。
【図１Ｃ】本発明による１つの受信機の機能ブロック図の一部である。
【図１Ｄ】本発明による１つの受信機の機能ブロック図の一部である。
【図２】相関図上の原疑似距離を示す。
【図３】カルマン・フィルタのブロック図である。[0001]
(Technical field)
The present invention relates to a radio receiver, and more particularly to a receiver for a wide area position measurement system used in a NAVSTAR wide area position measurement system (GPS).
[0002]
(Background technology)
The NAVSTAR wide area location system is a radio navigation and time transfer system based on a wide area satellite and uses a collection of US Department of Defense satellites. This assembly has at least 21 satellites located in multiple orbits at a predetermined distance on the earth, substantiallyZAt these locations, at least four satellites always exist on the horizon. There is one satellite control base that controls, among other things, the high-precision clocks in the satellites, synchronizes these clocks, confirms the orbits, ie the satellites, and orbit information to retransmit to the user. Upload to the satellite.
[0003]
Each satellite in this aggregate transmits two unique direct sequence spread spectrum messages in phase orthogonal to each of the two L-band frequencies. The receiver described here only processes C / A (coarse / acquisition) code messages on the L1 (1575.42 MHz) carrier, but the principle discussed here is the P (fine) code on L1 and L2. It also applies to signals. This direct sequence code signal has a frequency of 1.023 Mchips / second and a code repetition reference rate of 1 KHz and modulates the carrier wave in a binary transfer shift key (BPSK) scheme. This spread signal is further BPSK modulated by a 50 bps data signal. This data enables the receiver to measure the distance between the receiver and the satellite, that is, the standardization of the spacecraft's orbit (position estimation table) and the information based on the high-precision clock. Have data to do. These distance measurements are known as pseudoranges because the receiver clock is always offset from the satellite clock. To perform position resolution, four pseudo-range measurements are needed to resolve the four variables X, Y, Z and the local clock offset. To further improve the accuracy of position resolution, for satellite signal variables such as time of day (several measurements from the same satellite), wider satellite combinations, or wider phase and phase velocity, Further measurements can be made.
[0004]
A location reporting system using NAVSTAR GPS satellite signals is described in co-pending Australian Patent Application No. 63939/90.
[0005]
Many factors have influenced the basic design of the GPS receiver of the present invention. In addition to specific customer requirements, several design requirements have been identified.
[0006]
First, the receiver design must be flexible enough to be used in many applications. This means a modular design in which the individual modules are compatible. Furthermore, the ratio of receiver functionality to be provided by the software must be optimized by software written in modular form, preferably in advanced languages.
[0007]
Second, receiver designIsMust be suitable for a variety of potential environments. This means using appropriate margins and packages for hardware design. More important is the signal processing capability to operate under various dynamic conditions such as fluctuations as the vehicle moves or satellite signal vibration and intermittent blocking.
[0008]
Third, high performance, low cost receivers mean a reduction in manufacturing costs, which maximizes the functionality provided by the software or uses high density packaging technology.ZThis can be realized by reducing the size of the hardware. Furthermore, the receiver is designed to use a C / A code that is different from the more complex P code.
[0009]
Fourth, even when a C / A code is used, when used in a differential mode, the receiver can be made almost as accurate as a P-code receiver. This is a systematic correction of the pseudorange error for the receiver, so positioning is accurate.TheRequires a base base. The receiver must measure pseudoranges (and other parameters) with the resolution and accuracy required for accurate interpretation, so this measurement error and resolution is less than the error on the system being removed. Must be much smallerThisI must.
[0010]
Fifth, for portability and convenience, receivers should be designed to minimize weight and volume.ThisI must. Furthermore, since many proposed applications are to be carried remotely and used on battery power, power consumption must be minimized. Here, high-density mounting technology becomes important again.
[0011]
A known closed loop GPS receiver uses two types of loops. That is, these are two code lock loops that extract delay accumulation (pseudorange) of codes and phase lock loops (generally Costas loops) that extract data. The Costas loop can also be used to perform on-the-fly measurements of carrier parameters such as phase and phase velocity. In the code lock loop, the received signal is despread using a replica of the satellite code and the datasoThis despread signal is demodulated to be copyless using a modulated carrier replica. In order to match this replica with the received signal, the resulting energy is balanced between the early and late channels. In the Costas loop, this data is demodulated by coherent demodulation.
[0012]
This phase-lock approach only approximates optimality. Another problem is that in situations where the signal-to-noise ratio (SNR) is low, such as where there is significant noise, vibration, or jamming, the lock is lost or the period is likely to jump. The code lock loop also suffers performance degradation, but this is not critical to the overall receiver performance. This problem is natural in dynamic conditions. However, even under less dynamic conditions, such as with high precision exploration, the possibility of a cycle skipping is significant because the associated measurements take time.
[0013]
Other types of GPS receivers use an open loop estimator of signal parameters. In particular, a general correlation receiver can be used to make the signal parameter estimators most approximate. The estimate is obtained by selecting the processed signal parameters to perform position resolution.
[0014]
The dynamic performance of an open loop design is superior to that of a closed loop design because the former is not subject to phenomena such as loss of synchronization or loss of carrier lock. It is. This allows open loop receiver designers to be more flexible in changing parameters to meet changing application requirements. Furthermore, open loop design can be performed more cost-effectively because the basic processing is modular and there are fewer constraints on the designer in terms of acceptable processing delay.
[0015]
A receiver for a highly dynamic wide area location system is described in US Pat. No. 4,578,678 to WJ Hard. Unlike a hard receiver, the receiver of the present invention operates in phase (coherently). As a result, the SNR can be greatly improved. Furthermore, without a phase synchronization approach, the phase parameter of the carrier cannot be measured. This increases the number of parameters usedDoThis can lead to a significant improvement in accuracy. Although the hardware used P-codes, the receiver of the present invention uses C / A codes, which greatly simplifies the hardware and processing software. The hardware design depends on the specific hardware and processing software depending on the application. Since the processing is executed by software, the receiver of the present invention is a software application of the receiver.LeBy simply changing the golism, more flexibility is offered for many different uses. Another advantage to the hardware is the control that the processor has over the frequency of the local code generator, which allows the optimization of the measurement values constituted by the signal parameters.
[0016]
The GPS receiver according to the present invention produces code delay, carrier frequency, carrier phase, carrier phase acceleration, data values and data delay estimates. The receiver makes these estimates using an open loop correlation technique. By using multiple optimum parameter estimates, the accuracy of this receiver is high. Because the receiver is inherently open loop, the receiver is relatively resistant to large vibrations or very dynamic environments and can operate reliably even if the signal is intermittently blocked. Furthermore, since most of the signal processing is performed in software, the receiver is inexpensive and versatile.
[0017]
(Disclosure of the Invention)
Accordingly, the present invention comprises a GPS receiver for use in a satellite-based NAVSTAR wide area radio navigation and time transfer system, where the transmission from each of the plurality of satellites comprises a carrier modulated in code and data. And a receiver for receiving a C / A code signal transmitted from at least four satellites, a code delay, a carrier frequency, a carrier phase, a carrier wave Phase ofOKConsists of means for processing a signal having an open loop estimator of signal parameters for phase synchronization estimation of a plurality of signal parameters including speed, data value and data delay.
[0018]
This receiver incorporating the principles of the present invention can be designed for many different applications with various environmental and dynamic conditions.KiThe To illustrate the invention in greater detail, an embodiment of the invention will be described with reference to the accompanying drawings.
[0019]
(Best Mode for Carrying Out the Invention)
To facilitate the estimation of signal parameters and eliminate errors on the system, three of the parameters are not estimated directly, but as differences between selected satellites. Rather than making four estimates for the four pseudoranges, the shortest pseudorange is assigned to the nominal value zero, and therefore values above this zero value are assigned to other pseudoranges. This has been found not to cause errors in position resolution because one of the variables required for resolution is the clock offset. In effect, this estimation method allows the clock offset value to be set to the shortest pseudorange value. Not only does this approach eliminate errors common to each channel, but the number of bits needed to represent the pseudorange1/4 or moreThis reduces the size of the message that needs to be sent to the base base when using differential positioning.DecreaseCan be reduced. Differences are also estimated for phase, phase velocity (frequency) and phase acceleration parameters. This eliminates errors common to estimates from different satellites, such as errors due to reference oscillator phase variations.
[0020]
hardware
1A to 1DThese show the architecture of the receiver of a present Example. These functional blocks are hardware(10, 12, 14, 16, 20, 22, 24, 26, 28, 30)WhendigitalSoftware executed by the signal processor(32, 34, 36, 38, 40, 42, 44, 46, 48, 50, 52, 54, 56, 58, 60, 62, 64)It is divided into. Hardware amplified(10, 12)Down conversion(14, 26),filter(12, 28, 30)Perform functions, these are all “normal” functions that are performed in hardware. This down conversion is a “pseudo-signal stage”, ie a carrier waveButDownconverted to a frequency slightly higher than DC (relative to carrier frequency and signal bandwidth). This offset reduces the self-detection component and the local oscillator feed-through component.At 28 and 30It becomes possible to filter. Since this offset is very small compared to the signal bandwidth, it is necessary to generate an in-phase quadrature component.
[0021]
In this example, code generation and mixing isIn 20 and 22It is executed by hardware under the control of the main processor.In 26Further down conversion is also performed using a “composite” local oscillator with clearly defined limits. The frequency of the local oscillator is controlled by the processor, allowing drift and a predictable Doppler shift, thereby extending the effective range of carrier frequencies that the receiver can handle. MixingThe received signal at 16Clearly set the limit, ie, binary “1” if positive, “0” if negative,In 22Performed digitally in binary.In 16Setting the limits clearly results in a loss of SNR, but this simplification possible in hardware more than compensates for the disadvantages.
[0022]
In 22As a result of mixing the codes, the bandwidth of the signal will beAbout twice, ie 2MHzbecome. This signal isIn 24Sampled at 5 MHz, thenIn 28Filtered with a low pass filterThe9765Downsampled to Hz, this frequency is low enough to process using software.Shown by this 28The filter / downsampler is implemented as an integration and dump that integrates (counts) a number of input samples to give one output sample. The integral and dump filters have a sinc-function frequency response, which is sampledringHas first zero at frequency. This means that there are many pass bandwidths at frequencies higher than the Nyquist frequency. Therefore, the signal and noise frequency components in this region pass as noise due to aliasing. This problem can be easily solved by integrating twice as many input samples for the same output sampling rate. This means that a given output sample is the sum of the two previous integrals and dumps. This second sum is the anti-aliasing filter30Known as and implemented in hardware.
[0023]
Main processor
Up to this point, the functional block of the receiver is either operating continuously, or if sampled, each sample has no special meaning with respect to the position of the receiver by time. . In the main processor, all algorithms are based around a 25 ms time unit known as the “integration period”.1A to 1DAs shown, one algorithm is executed in each integration period and the other algorithm is executed once every 3 to 54 integration periods.
[0024]
Also, some of the main processor algorithms are executed according to the acquisition status of the receiver, i.e. whether the satellite is being acquired. The input signal is32,Transform to frequency domain by using Fast Fourier Transform (FFT). Each bin in the FFT isIn 36Check the size. If any bin does not exceed the threshold, the satellite signal is considered not captured and the main processor waits for the next integration period. If the threshold is exceeded, this bin where the maximum value occurs is, 34Narrowband Infinite Impulse Response (IIR) low pass filter down converted to zero Hz bin38Apply to this time domain. The signal at this stage is a carrier wave spread by 50 Hz data.
[0025]
At 40Of this signalUnwrappedFind the phaseIn 42Differentiate over the integration period. by this,In 44Data bit transition values (180 ° phase transition values) can be detected. Once the location of the transition value string is determined,At 60Data can be decrypted. Looking for frame information repeated by this operation and obtaining this information, the polarity ambiguity is removed and the remaining data can be decoded. The purpose of this decoding is to set the transmission time of the transition value of the data and to extract the position estimation data. This data indicates the orbit of the satellite with high accuracy.
[0026]
In 44Based on these detected transition values,46,Signal,Replication of detected dataMultiplication withDespread by. The optimal detection method can be ensured by performing some of these despreading operations before and after the detected transition value and picking up the peak of the correlation function against the delay parameter of the data. If the signal is despread with respect to the data, leaving only the nominal carrier,In 48Measure the phase again. A linear fit is performed on this phase function. by this,46,Produces a least-squares estimate of the phase velocity over the integration period. Next, by linear fittingObtainedLineReal axisUsing this conversion converted toIn 52The phase of this signal is rotated.In 54Of this rotated signalReal valueThe sum of,In effect, this is the phase synchronization correlation of this signal. Each integration periodInOne of these correlation values is created. This set of thresholds on the correlation value also indicates that the satellite has not been acquired.
[0027]
Estimate code delay to measure pseudorangeTheIt must be made. In this receiver, pseudoranges are not directly measured, but are measured as differences between satellites. Zero value for one satelliteTheAssign to other satellites the difference between this pseudorange and the pseudorange in question. Measuring the difference simplifies the receiver in that no high resolution clock is required and simplifies communication with the base base.
[0028]
Code delay values are code generators in three different "phases" for three consecutive integration periods20Is estimated. These three phases “lock” (ie, align the local and satellite codes) with a portion of the chip early and a portion of the chip late. The correlation function can be reconstructed from the phases of these three codes. The best way to reconstruct the correlation function and pick up the peaks isIn 54Correlation function measured by sampling at 3 pointsIn 56Correlating with the expected correlation function, which can be measured preferentially. By using this double correlation method, an estimate of the original pseudorange is generated. The use of this second correlation is necessary to optimize the application consisting of three integration times that contribute to the measurement. Each of the three correlation sample values is obtained by observing the signal for only one integration periodIn 54Although generated, the total measurement time is three integration periods. The noise on the sample is not related between the sample values. Therefore, this sampled correlation curve is semi-optimal. By correlating this correlation curve with the expected correlation function, the correlation signal-to-noise ratio is maximized, thereby restoring the optimality of the process and allowing interpolation of these sample values.
[0029]
Over a period much longer than the integration period (this length is set by the signal detection requirements) to meet the requirement to accurately measure pseudorange differences, phase differences, frequency differences and phase acceleration differences The signal must be observed. In other words, generate a more accurate estimateTheTherefore, these original estimates must be combined in some optimal way. The method chosen to combine these estimates is the Kalman filter62Is to use. This filter62Inputs to pseudorange, frequency and phaseofThe estimated difference and the pseudo distance from the basecorrectionValue. This filter62Model receiver position, satellite and receiver movement, and receiver oscillator drift. From these models and inputs, this filter not only generates a smooth but uncorrected pseudorange for transmission to the base station, but also generates an estimate of the receiver position for the user. These estimates are64,Provided to the user with a standardized reference system that requires geometric transformation.
[0030]
correlation
The closest (ML) estimator is a correlation detector. The signal parameters in question are code value, signal delay (measured using code delay and data delay), carrier frequency, carrier phase, and data value. The receiver must effectively perform a 6-dimensional correlation and pick up its peaks to obtain an optimal estimate of all parameters in order to obtain an optimal position resolution.
[0031]
Two other parameters that appear in the estimation process are local oscillator drift and code generator frequency. Oscillator drift is important because all time measurements made in the receiver are made with reference to one oscillator, which is used for the local oscillator and the sampling clock. To do. Kalman filter62To generate an estimate of the oscillator frequency and use thisTheCode generator20The frequency setting can be corrected. Code generator20By correcting the frequency, the satellite code can be tracked precisely, the correlation shape and the correlation value can be optimized, and the SNR can be maximized. This is not estimated as an independent parameter (ie, no correlation is performed on the corrected frequency) because it is directly proportional to the already estimated frequency of the satellite signal carrier.
[0032]
Code value
The choice of code value is simply the satellite code to be used for each channel of the receiver.ofRequires selection. Each satellite has an epoch length code of 1023 chips, and the satellite repeats this code at an epoch rate of 1 KHz. In the receiver,20Each satellite channel generates a copy of the satellite code from which data is to be received.22To this locally generated code,Multiply the received satellite signal. If the local code is delayed so that the codes match, the received signal is despread from the 2 MHz spread spectrum bandwidth to the 100 MHz data bandwidth.
[0033]
Estimating data values
The data value is obtained by searching for the transition value of the data. The correct polarity of the data isAt 60Set by data decryption algorithm. For the purpose of signal processing, this data value is estimated as an inverted or non-inverted value of the actual data, and is determined to be “1” or “0” later.42,The transition value is set in time by filtering the phase of the carrier modulated with data and matching the filter with the transition value.In 42Matched filter output peaks occur in the region of data transition. Filter matching is the best way to set transition values and is accurate within a small number of samples. A code epoch can be used to unambiguously set the transition value and give an initial estimate from the matched filter. The matched filter sets the transition value in a small number of samples (ie within a few hundred microseconds), the data transition value always corresponds to the code epoch transition value, which occurs every 1 ms, If the position of the code epoch boundary within the integration period is known, the transition value can be explicitly set. This is the method used in this receiver.
[0034]
When data transition values are set and decoded, the information in this data is, 56 and 58Contributes to other measurements. The timing information in this data is used to set the time of the week and allow the receiver to calculate the end transmission time used to measure the pseudorange. Using the data in the position estimation table, the satellite position is accurately determined and the transmission time is given.
[0035]
Estimation of phase, phase velocity (frequency) and phase acceleration
An estimate of the frequency of each satellite channel was used to detect the signal synchronouslyIn 14, 26 and 34It is simply the sum of all down conversion values. Nominally there are four such down conversion values, one in analog hardware (by LO)Of 14And one in digital hardwareOf 26And two in the main processor (one in the IIR filter)34 before 38And one isIn 52Both are in the time domain) due to phase rotation. Each of these conversion values is nominally smaller than the previous conversion value. Each measured value is proportional to the LO frequency. This is because the down conversion frequency performed in the hardware and processor is all due to the sampling frequency, and this sampling frequency is taken from the same source as LO. In other words, LO drift causes a multiplication error in this measurement. However, by estimating the frequency difference, the first LO frequency offset is removed and the drift only increases the down conversion performed in software.
[0036]
The optimality from the frequency estimate isIn 34 and 52Only relevant to the last two down conversion values,In 14 and 26The first two are effectively “constants”. The FFT is correlated with frequency, so estimating the frequency for the first down-conversion within the main processor with the FFT gives an ML estimate.40, 46 and 48The phase processing uses averaging over frequency and phase estimates, which can be shown to be an optimal interpolation of complex FFT estimates.
[0037]
If the phase is measured at the boundary of the integration period, that measurement is also optimal. The phase contribution due to the first LO is eliminated by using the phase difference.In 26Down conversion in digital hardwareIn 34Any FFT down-conversion contributes to bringing the signal to zero phase at the boundary of the integration period. This is due to these down conversion “oscillators” having an integrated number of periods within one integration period and initialized to zero. This means that the estimated value of the phase is an average value estimated during the phase processing period. In order to remove phase ambiguity, this phase estimate must be “no wrap” with respect to 2.
[0038]
The phase acceleration is not optimal because it does not use all available information. This uses only early and late integration periods, and generally these integration periods both have a lower SNR than those that adhere to time.
[0039]
Delay estimation (pseudo distance)
The delay estimate is more complex than the other estimates, but is the most fundamental in determining the receiver location. As explained earlier, only the pseudorange differences are estimated. This estimate consists of a code delay (which is 1 ms ambiguity, but can be easily estimated because the code is known) and a data delay (which is not ambiguous but not known) Requires input from both.56 and 58,Tag time to pseudorange measurements using data transitions that occur on each channel. This sets data transitions within one sample, as described in Estimating Data Values. Data transitions always coincide with code epoch transitions. Interpolation between samples is56,The "phase" of the code at the start and end of the integration period,Sub chipThis is possible by sampling with accuracy. The frequency of the received code to such an extent that the local code frequency can be estimatedThe"Lock"56,Interpolate points between samples where transitions occur. If the local code exactly matches the received code (with delay) and there is no noise in the system, this interpolated pseudorange measurement is perfectly correct.
[0040]
However, the requirement for both maintaining code lock and making an accurate estimate despite system noise requires two additional processing steps. The first of these steps isIn 56Picking up a correlation peak for the delay parameter, the second step comprises:In 56Including combining a series of measurements taken over an observation period to ensure that the requirements for accuracy are met.
[0041]
The pseudorange measurement created using the method described above has been referred to as the “original pseudorange”. These estimates are not optimal because there is no relationship between these estimates and the correlation in the delay parameter. This is handled by making three such estimates within three consecutive integration periods, and making each estimate for a different “phase” of the code. One of the integrations is performed by a code set to an “optimal estimate” of the phase of the code (ie, approximated so that it can be estimated to be “locked”). The other two are done in the early point part of the chip and the late part of the chip,FIG.A set of measurements is given as These three points must be used to reconstruct the shape of the correlation function in order to find the exact peak of the correlation and thus correct the pseudorange measurements made with the “best” estimate. The best way to reconstruct is56,These three points are the expected correlation shape (As shown in FIG.). This estimate of pseudorange is called the “interpolated pseudorange” and is the best estimate because it uses the correlation to create the estimate (two estimates-the shape of the correlation Estimates and correlation peak estimates-Since, in fact, 2 times). However, even though each estimate is optimal, they are still not accurate enough. Some of these interpolated pseudoranges (usually at least 18)62To produce a final estimate of the pseudorange and position. Kalman filter62The estimated phase and frequency values input to the signal contribute to improvement of the estimated pseudorange.
[0042]
Kalman filter
Kalman filter62The function of is to track the “state” of a linear dynamic system by processing the observations in an iterative least squares method. The state of this system is described by a vector of system variables, which must be related to each other by a set of known linear equations. In general, a state vector is composed of a set of one or more derivatives, so the equations associated with these vectors are differential equations that represent the dynamics of the system. Observed values are made by measuring observable values, and these observable values must be known linear functions of the state variables.
[0043]
FIG.The Kalman filter62Shows the operation. The observation vector Y is composed of three pseudo-range differences and three phase differences. The state vector is X and is composed of the receiver position, velocity and acceleration on three coordinate axes, and the rate of change of the local clock offset. “Hats” indicates estimated values, ie, estimates.
[0044]
Transition matrix φ

To obtain an estimate of the state vector for the next time interval. This loop is known as a state transition loop. Observations are made every N transitions,

Is corrected. Thus, during the observation, the transition matrix models a predictable dynamic behavior of the receiver.
[0045]
Expected observation vector

Is compared to the observation vector Y and the resulting prediction error vector is processed to produce a correction value. The latter is multiplied by a so-called karma gain matrix K. The latter is calculated by a probabilistic model technique including a state covariance matrix P, an observation error covariance matrix R, a system noise covariance matrix Q, and an observation matrix M. This observation matrix associates measurement vectors with state vectors.
[0046]
There are descriptors on the matrices that form the models (eg, φ, Q, and M), which means that these matrices are time invariants and must be updated regularly. Show. These matrices depend on the satellite and receiver positions. Furthermore, Q and R are filters62Can be estimated from the observations by various methods that allow the receiver to adapt to changes in the dynamic behavior of the receiver. filter62Another method that can be applied is by multiplying a scalar coefficient related to the locus of the estimated value of the prediction error. Specifically, the scalar coefficient is 1 + α, whereαIs a matrix

Linearly related to the subtrajectory of
[0047]
Kalman filter62The observations related to are made by input from a pseudorange interpolation algorithm and input from a frequency and phase estimator. After a predetermined number of observations have been made, the position estimate can be output directly to the user from the expected state vector. Similarly, an estimate of the pseudorange difference that is to be differentiated and transmitted to the base station can be extracted directly from the expected observation vector. If the pseudorange correction value is available from the base station,FIG.Filter, as shown in62These values can be used to generate a more accurate position by supplying to
[0048]
Positioning algorithm
The positioning algorithm used in this receiver is the Kalman filter.62Executed within.
[0049]
It is emphasized that the present invention is not limited to the embodiments described here. Many modifications and variations can be made within the broad concepts described. The receiver design can be improved in several ways.
[0050]
The generic concept described here relates to a receiver having four or more channels, known as a “multi-channel” receiver.
[0051]
With little effort, this concept is a 3 channelshipIt can be easily applied to uses such as a marine receiver or a one-channel multiplex receiver. In the case of a one-channel receiver, the important difference is not the difference but the absolute measurement values such as pseudorange, phase, etc. are required. Other modifications and variations are intended to accommodate specific applications and special environmental conditions.
[Brief description of the drawings]
FIG. 1 is a block diagram showing the arrangement of FIGS. 1A to 1D.
FIG. 1A is part of a functional block diagram of one receiver according to the present invention.
FIG. 1B is part of a functional block diagram of one receiver according to the present invention.
FIG. 1C is part of a functional block diagram of one receiver according to the present invention.
FIG. 1D is a part of a functional block diagram of one receiver according to the present invention.
[Figure 2]The original pseudorange on the correlation diagram is shown.
[Fig. 3]It is a block diagram of a Kalman filter.

Claims (24)

A GPS receiver used for satellite-based radio navigation that receives a plurality of satellite signals including a phase-inverted modulated carrier signal representing a plurality of satellite identification codes and data from a plurality of satellites,
Means (12) for receiving and amplifying the plurality of satellite signals via an antenna;
Code generating means (20) for generating any one of the plurality of satellite specific codes, wherein the code generating means (20) generates a code such that a code delay of the generated code can be controlled with subchip accuracy;
Code mixing / down converting (code mixing / down converting) means (14) for multiplying the amplified satellite signal by the generated code (22) and down-converting the carrier frequency (14, 26) to produce a digital signal , 22, 26), wherein the digital signal has a down-converted carrier frequency and is multiplied by the generated code, so that the code delay of the generated code is reduced. The bandwidth of the portion of the digital signal resulting from the satellite signal having the same code is despread when the code delay of the satellite signal having the same code is the same or nearly the same. Code mixing / downward conversion means (14, 22, 26);
Means (32, 34, 36, 38, 40, 42, 44, 46, 48, 52, 54, 56, 58, 60) for calculating the signal parameters of the satellite signal including the code delay;
Means (62, 64) for providing information about the position of the antenna using signal parameters from at least four of the plurality of satellite signals;
In a GPS receiver comprising: a method for determining a code delay of the satellite signals having the same code;
(1) generating a plurality of the digital signals, each of the digital signals corresponding to a different code delay of the code generated from the code generating means ( 20, 22, 24, 26, 28, 30),
(2) A plurality of correlation values corresponding to each digital signal, each correlation value being generated using information on both the carrier phase and the carrier frequency and indicating the degree of code demodulation of the digital signal Calculating correlation values of (32, 34, 36, 38, 40, 42, 44, 46, 48, 50, 52, 54);
(3) calculating an optimal code delay of the generated code from the plurality of correlation values to provide a code delay of the satellite signal;
Calculating the plurality of correlation values comprises :
(A) step of calculating over an integration period at least one composite frequency spectrum of the digital signal (32),
(B) finding the position of the largest bin in the frequency spectrum (36) ;
And (c), characterized in that it comprises a step (36,46,48) for determining the carrier frequency offset and carrier phase. A GPS receiver used for satellite-based radio navigation that receives a plurality of satellite signals including a phase-inverted modulated carrier signal representing a plurality of satellite identification codes and data from a plurality of satellites,
Means (12) for receiving and amplifying the plurality of satellite signals via an antenna;
Code generating means (20) for generating any one of the plurality of satellite specific codes, wherein the code generating means (20) generates a code such that a code delay of the generated code can be controlled with subchip accuracy;
Code mixing / down converting (code mixing / down converting) means (14) for multiplying the amplified satellite signal by the generated code (22) and down-converting the carrier frequency (14, 26) to produce a digital signal , 22, 26), wherein the digital signal has a down-converted carrier frequency and is multiplied by the generated code, so that the code delay of the generated code is reduced. The bandwidth of the portion of the digital signal resulting from the satellite signal having the same code is despread when the code delay of the satellite signal having the same code is the same or nearly the same. Code mixing / downward conversion means (14, 22, 26);
Means (32, 34, 36, 38, 40, 42, 44, 46, 48, 52, 54, 56, 58, 60) for calculating the signal parameters of the satellite signal including the code delay;
Means (62, 64) for providing information about the position of the antenna using signal parameters from at least four of the plurality of satellite signals;
In a GPS receiver comprising: a method for determining a code delay of the satellite signals having the same code;
(1) generating a plurality of the digital signals, each of the digital signals corresponding to a different code delay of the code generated from the code generating means ( 20, 22, 24, 26, 28, 30),
(2) A plurality of correlation values corresponding to each digital signal, each correlation value being generated using information on both the carrier phase and the carrier frequency and indicating the degree of code demodulation of the digital signal Calculating correlation values of (32, 34, 36, 38, 40, 42, 44, 46, 48, 50, 52, 54);
(3) calculating an optimal code delay of the generated code from the plurality of correlation values to provide a code delay of the satellite signal;
Calculating the plurality of correlation values comprises :
(A) step of calculating over an integration period at least one composite frequency spectrum of the digital signal (32),
(B) finding the position of the largest bin in the frequency spectrum (36) ;
And (c), characterized in that it comprises a said maximum step in the time domain according to the position of the bottle by running further downconverted to respective digital signals make further down converted signal with a residual carrier frequency (34). Wherein the plurality of calculating a correlation value The method of claim 1 including the step (46) to remove it if there is any data transitions said at least one of said digital signal. Wherein the plurality of calculating a correlation value The method of claim 2 including the step (46) to remove it if there is any data transitions from said at least one of said further down-converted signal. The method of claim 1, wherein the plurality of digital signals are generated sequentially (20) . A GPS receiver used for satellite-based radio navigation that receives a plurality of satellite signals including a phase-inverted modulated carrier signal representing a plurality of satellite identification codes and data from a plurality of satellites ,
Means (12) for receiving and amplifying the plurality of satellite signals via an antenna;
Code generating means (20) for generating any one of the plurality of satellite specific codes , wherein the code generating means (20) generates a code such that a code delay of the generated code can be controlled with subchip accuracy ;
Code mixing / down converting (code mixing / down converting) means (14) for multiplying the amplified satellite signal by the generated code (22) and down-converting the carrier frequency (14, 26) to produce a digital signal , 22, 26), wherein the digital signal has a down-converted carrier frequency and is multiplied by the generated code, so that the code delay of the generated code is reduced. The bandwidth of the portion of the digital signal resulting from the satellite signal having the same code is despread when the code delay of the satellite signal having the same code is the same or nearly the same. Code mixing / downward conversion means (14, 22, 26);
Means (32, 34, 36, 38, 40, 42, 44, 46, 48, 52, 54, 56, 58, 60) for calculating the signal parameters of the satellite signal including the code delay;
And means (62, 64) that utilizes at least 4 Tsukara signal parameter of said plurality of satellite signals gives information about the position of the antenna,
In a GPS receiver comprising: a method for determining a code delay of the satellite signals having the same code;
(1) step of a plurality of said digital signal, for generating a plurality of said digital signal, such as different code delays codes each of said digital signal is generated from said code generation means corresponds ( 20, 22, 24, 26, 28, 30) ,
(2) A plurality of correlation values corresponding to each digital signal, each correlation value being generated using information on both the carrier phase and the carrier frequency and indicating the degree of code demodulation of the digital signal Calculating correlation values of (32, 34, 36, 38, 40, 42, 44, 46, 48, 50, 52, 54) ;
(3) the optimal code delay of the generated code calculated from said plurality of correlation values and a step (56) providing a code delay of the satellite signals,
Calculating the plurality of correlation values comprises :
(A) step of calculating over an integration period at least one composite frequency spectrum of the digital signal (32),
(B) the step of finding the location of the maximum bin in the frequency spectrum (36),
Method characterized by comprising the steps (34, 52) which in (c) the time domain by performing a further downconverted to respective digital signals make further down converted signal having a zero frequency and a zero phase. Said further down-conversion to zero frequency and a zero phase is performed in two stages (34 and 52), the first stage (34) includes a lower conversion to residual carrier frequency according to the position of the maximum frequency bin, and the The method of claim 6 , wherein calculating a plurality of correlation values further comprises estimating a residual carrier frequency (46) . The method of claim 7 , wherein calculating the plurality of correlation values further comprises estimating (48) a carrier phase offset of the down-converted signal. The digital signal and the further down-converted signal are represented by a phase vector having an in-phase component (I) and a quadrature component (Q), and the remaining carrier frequency measures the rotational speed of the phase vector of the further down-converted signal ( 46) The method of claim 7 estimated by: The method of claim 9 , wherein the rotational speed of the phase vector is measured using a linear fit (46) . Creating a further down-converted signal having zero frequency and zero phase, rotating the phase vector about the real axis according to the measured carrier frequency (46) and carrier phase offset (48) (52); And calculating a plurality of correlation values comprises:
(1) Detect (44) and remove (44) and remove data transitions, and (2) Aggregate the real values of the rotated phase vectors over an integration period to give their correlation values (54) ) to the method described <br/> to claim 6 which contains for each digital signal. Said code mixing and down-conversion means (14,22,26) is a down-conversion of the carrier frequency run in two stages (14, 26), a later stage of this two-stage (26) is the code mixture (22 ) is carried out after method of claim 1 degree down-conversion carried out in a later stage (26) which is adjusted according to the determined carrier frequency or maximum bin value (36). Method according to claim 1, wherein said code mixing and down-converting means (14, 22, 26) downsamples (28) the signal to a sampling rate suitable for further processing by computer software. Calculating a optimum code delay of the generated code from said plurality of correlation values (56), a plurality of correlation values, and predicted relationship between those correlation values and code delay of the satellite signal The method of claim 1 including comparing and determining a code delay for a satellite that best matches between the plurality of correlation values and their predicted relationships. A GPS receiver operating according to the method of claim 1. The signal parameters may include a pseudo-away difference in said at least four between the satellite signal, GPS receiver pseudo distance difference is actuated by the method of claim 1, calculated from the code delay of each of the satellite signals Machine. The signal parameters, GPS receiver that operates by the method according to claim 1, which contains the offset of carrier phase calculated from the digital signal. The signal parameters, GPS receiver that operates by the method of claim 1 including the acceleration of the carrier phase calculated from the digital signal. The signal parameters, GPS receiver that operates by the method of claim 1 including a carrier frequency calculated from the digital signal. The signal parameters include the estimated residual carrier frequency , the degree of further down-conversion in the first stage (34), and the degree of down-conversion performed by the code mixing and down-conversion means (14, 22, 26). A GPS receiver operating according to the method of claim 7 , including carrier frequencies calculated by summing (62) . The signal parameters, GPS receiver that operates by the method according to claim 9 comprising a carrier phase acceleration calculated by measuring the rate of change of the rotational speed of the phasor. Said means for utilizing the signal parameters, GPS receiver that operates by the method of claim 1 including a Kalman filter (62) for receiving said parameter. It said means for utilizing the signal parameters comprise a Kalman filter whose input (62) of the parameter, the signal parameter is (1) a carrier phase offset,
(2) Acceleration of carrier phase ,
(3) pseudo distance difference between the satellite signal obtained from the code delay of the satellite signals, and (4) GPS receiver that operates by the method of claim 6 including the carrier frequency. Further down conversion to the zero frequency and zero phase is performed in two stages (34 , 52) , the first stage (34) includes down converting to the remaining carrier frequency according to the maximum frequency bin, where the carrier frequency is 24. A GPS receiver according to claim 23 , wherein the GPS receiver is calculated using the degree of further down-conversion in the first stage and the residual carrier frequency.

Images