Positioning device

The positioning device enhances GNSS accuracy by converting real-value biases to integers using a reference station and multiple conversion units, addressing ambiguities and signal disruptions for stable and precise positioning.

JP7577410B2Active Publication Date: 2024-11-05JAPAN RADIO CO LTD
View PDF 5 Cites 0 Cited by

Patent Information

Application Number
JP2020176058
Authority / Receiving Office
JP · JP
Patent Type
Patents
Current Assignee / Owner
Filing Date
2020-10-20
Publication Date
2024-11-05
Estimated Expiration
2040-10-20

AI Technical Summary

Technical Problem

Conventional GNSS positioning systems face challenges in achieving highly accurate positioning due to ambiguities in carrier wave phase observations, particularly when encountering instantaneous interruptions and cycle slips, which hinder the conversion of real-value biases to integers, leading to inaccurate positioning results.

Method used

A positioning device that utilizes a reference station and receiver to calculate relative positions, employing a real number solution calculation unit, first and second integer conversion units, and integer solution variance adjustment to enhance the conversion of real-value biases to integers, thereby improving positioning accuracy.

Benefits of technology

The solution enables continuous and highly accurate positioning by increasing the number of integer-value biases used, stabilizing the positioning process even in environments with signal interruptions, and reducing the influence of unstable conditions on real-value bias updates.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 0007577410000002
    Figure 0007577410000002
  • Figure 0007577410000003
    Figure 0007577410000003
  • Figure 0007577410000004
    Figure 0007577410000004
Patent Text Reader

Abstract

To continuously obtain highly accurate positioning positions even in environments with many instantaneous breaks or cycle slips in ranging signals.SOLUTION: A positioning device for calculating a relative position of a GNSS receiver 1 relative to a reference station by receiving ranging signals transmitted from a plurality of satellites Sx by the reference station and the GNSS receiver 1 includes: a real number solution calculation unit 81 that calculates real number value biases by solving a positioning equation in which a three-dimensional position of the GNSS receiver 1 and the real number value biases of a carrier of the ranging signals are unknowns with carrier phases of the ranging signals as observation amounts; a first integer conversion unit 82 that generates a first subset and converts the same to integers by real number value biases that satisfy a predetermined condition among the calculated real number value biases; and a second integer conversion unit 83 that generates a second subset by adding, to the first subset, real number value biases whose absolute values of difference from the integer-converted integer value biases are smaller than a predetermined threshold among the real number value biases that were not subjected to integer conversion targets in the process of the first integer conversion unit 82, and converts the same to integers.SELECTED DRAWING: Figure 1
Need to check novelty before this filing date? Find Prior Art

Description

[Technical field]

[0001] The present invention relates to a positioning device, and more particularly to a technique for receiving distance measuring signals transmitted from satellites of a global positioning satellite system and performing relative positioning using observation data from an observation station and observation data from a reference station. [Background technology]

[0002] Global positioning satellite systems (GNSS: abbreviation for Global Navigation Satellite System), including GPS (short for Global Positioning System), Galileo, GLONASS (short for Global NAvigation Satellite System), BeiDou, and QZSS (short for Quasi-Zenith Satellite System), are systems that calculate the position of a receiving point using radio waves (specifically, ranging signals) transmitted from satellites. A positioning device that uses the global positioning satellite system to measure a position measures the position of the positioning device based on each signal transmitted from multiple satellites.

[0003] As a conventional technology for performing relative positioning using GNSS, a carrier phase relative positioning device is known which includes: means for receiving carrier phase information of a plurality of positioning satellites determined by a base station; means for receiving signals from the plurality of positioning satellites to determine each carrier phase, and determining a double phase difference of the carrier phase from the carrier phase and the carrier phases of the plurality of positioning satellites received from the base station, and locating a reception point from the double phase difference and the positions of the plurality of positioning satellites; and determination means for determining the presence or absence of a cycle slip occurring in the double phase difference based on the difference between the height of the reception point and the height of the reception point determined without using received signals from the positioning satellites (see Patent Document 1). [Prior art documents] [Patent documents]

[0004] [Patent Document 1] JP 2001-013237 A Summary of the Invention [Problem to be solved by the invention]

[0005] In the positioning using GNSS, the carrier phase equivalent to the pseudo distance from the satellite to the positioning device is observed, and the observed carrier phase (called "carrier observable") is used to obtain a highly accurate positioning position. However, while the phase part of the carrier can be accurately observed, the number of wavelengths during radio wave propagation (in other words, "bias" equivalent to the number of wavelengths) cannot be accurately observed. For this reason, ambiguity in the number of wavelengths always exists in the carrier observable. By resolving the ambiguity (in other words, accurately estimating the number of wavelengths during radio wave propagation), a highly accurate positioning position can be obtained. To resolve the ambiguity, the ambiguity is solved as an unknown quantity in the positioning equation to obtain a real-valued bias, and further, a subset of the real-valued bias is prepared, and the real-valued bias group of the subset is converted into integers using the integer least squares method to obtain an integer-valued bias. Then, a highly accurate positioning position is obtained using the obtained integer-valued bias.

[0006] Here, the positioning position obtained using the integer bias is called a "fix solution," and the positioning position obtained using the real bias without the integer bias is called a "float solution." The positioning accuracy of the fix solution is better than that of the float solution, so the process of converting the real bias to an integer (in other words, the process of obtaining the integer bias) is very important.

[0007] However, the number of integer biases obtained by the integer conversion process of real-valued biases is small. There is a problem that, if the small integer bias is not used, the positioning position obtained with the small integer bias has a large error. Specifically, for example, when a moving user (in other words, a GNSS receiver) receives ranging signals transmitted from satellites A and B to obtain a positioning position, the ranging signals transmitted from satellites A can be received stably, while the ranging signals transmitted from satellites B are likely to be interrupted or have cycle slips due to utility poles or electric wires. In this case, the observation error of satellites B is large, and errors are also superimposed on the estimated real value bias, so the real value bias of satellites B cannot be converted into an integer, that is, the integer value bias cannot be obtained. For this reason, the integer value bias is obtained only from the ranging signals transmitted from satellites A. In this case, the error superimposed on the integer value bias is not zero, so errors are easily superimposed on the positioning position obtained with the small integer value bias, and the accuracy of the obtained three-dimensional positioning position deteriorates according to the spatial bias of the small integer value bias.

[0008] In addition, the ranging signals transmitted from the B satellites are subject to instantaneous interruptions and cycle slips due to utility poles and power lines, and the continuous reception of the ranging signals is temporarily interrupted, causing the received signals to become momentarily unstable, resulting in larger observation errors, and it takes time for the observation errors to converge again. In situations where the observation errors are large, large errors are also superimposed on the real value bias estimated based on the observation amount, making it impossible to obtain the correct integer value bias. In order to obtain the correct integer value bias, it is necessary not to convert the real value bias of the B satellites to integers until the observation errors converge. In this way, in situations where instantaneous interruptions and cycle slips are likely to occur, the observation errors are difficult to converge, so the number of integer value biases obtained by successfully converting the real value bias to integers remains small, resulting in a problem of continuously deteriorating positioning accuracy. Furthermore, while the user (in other words, the GNSS receiver) is moving, the conditions are constantly changing, such as conditions in which the ranging signal can be received stably and conditions in which momentary interruptions and cycle slips are likely to occur. As a result, there are many cases in which the observation error that has converged and become small increases again. As a result, it takes time for the observation error to converge completely, so the real-valued bias that can be converted to an integer decreases, and it is not possible to continuously obtain a stable positioning position.

[0009] Therefore, an object of the present invention is to provide a positioning device capable of determining a highly accurate positioning position, for example, capable of continuously determining a highly accurate positioning position even in an environment where there are frequent interruptions or cycle slips in the ranging signal. [Means for solving the problem]

[0010] In order to solve the above problem, the present invention provides a positioning device in which a reference station and a receiver receive ranging signals transmitted from a plurality of satellites and calculate a relative position of the receiver with respect to the reference station, the positioning device including a real solution calculation unit that calculates the real-valued bias by solving a positioning equation in which a carrier phase of the ranging signal is an observable and a three-dimensional position of the receiver and a real-valued bias of a carrier wave of the ranging signal are unknowns; a first integer conversion unit that generates a first subset by the real-valued biases that satisfy a predetermined condition among the calculated real-valued biases and converts them into integers; and a real solution calculation unit that converts the real-valued biases that are not subject to integer conversion processing in the first integer conversion unit among the calculated real-valued biases by generating a first subset by the real-valued biases that satisfy a predetermined condition among the calculated real-valued biases. 、 The first integer conversion unit The real-valued bias that is the subject of the integer conversion process. Difference with Decimal part of a second integer conversion unit that generates a second subset by adding the real-valued bias, the absolute value of which is smaller than a predetermined threshold, to the first subset and converts the second subset into an integer; death , When the first subset includes all of the real-valued biases or when the integerization of the second subset is not successful, the set of integer-valued biases output from the first integerization unit is used to correct the three-dimensional position of the receiver calculated in the real solution calculation unit, and the relative position of the receiver with respect to the reference station is calculated; when the first subset does not include all of the real-valued biases and the integerization of the second subset is successful, the set of integer-valued biases output from the second integerization unit is used to correct the three-dimensional position of the receiver calculated in the real solution calculation unit, and the relative position of the receiver with respect to the reference station is calculated. The positioning device is characterized by the above.

[0011] The invention according to claim 2 provides the positioning device according to claim 1, further comprising: a first integer conversion unit converting the first integer into an integer number based on a success or failure of the first integer conversion unit and a success or failure of the second integer conversion unit converting the second integer into an integer number. The above The method further comprises an integer solution variance adjustment unit that adjusts the variance of an integer value bias, and a real solution update unit that updates the real value bias calculated by the real solution calculation unit using the adjusted variance of the integer value bias.

[0012] The invention as recited in claim 3 is the positioning device as recited in claim 1 or 2, characterized in that the positioning equation is solved by applying a Kalman filter. Effect of the Invention

[0013] According to the invention described in claim 1, a second subset is generated by adding real-valued biases that satisfy a predetermined condition among the real-valued biases that were not converted into integers in the previous integer conversion process to the first subset, and then the real-valued biases are converted into integers. This makes it possible to increase the number of integer-valued biases, and by performing positioning using a larger number of integer-valued biases, it becomes possible to obtain a highly accurate positioning position.

[0014] According to the invention described in claim 2, the variance of the integer value bias is adjusted based on the integer conversion situation, and the real value bias is updated using the adjusted variance of the integer value bias. Therefore, in an unstable situation where interruptions or cycle slips are likely to occur, the variance of the integer value bias is adjusted to be larger, so that the real value bias can be updated with an appropriate variance of the integer value bias. This reduces the effect that the integer value bias with a large variance has on the update of the real value bias, and makes it possible to suppress the update error of the real value bias and continuously obtain a stable, highly accurate positioning position even if the surrounding environment changes significantly.

[0015] According to the invention as defined in claim 3, it is possible to appropriately solve the positioning equation. [Brief description of the drawings]

[0016] [Figure 1] 1 is a functional block diagram showing a schematic configuration of a GNSS receiver including a positioning device according to an embodiment of the present invention. [Diagram 2] 2 is a flow diagram showing a processing procedure in the GNSS receiver of FIG. 1. [Diagram 3] 3 is a flowchart showing a processing procedure in a second integer conversion unit, and is a flowchart showing details of the processing in step S9 in the flowchart of FIG. 2. FIG. [Figure 4] 10A and 10B are diagrams illustrating situations related to the conversion of real-valued biases into integers by a positioning device according to an embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0017] The present invention will be described below based on the illustrated embodiment. Fig. 1 is a functional block diagram showing a schematic configuration of a GNSS receiver 1 including a positioning device 8 according to an embodiment of the present invention. Fig. 2 is a flow diagram showing a processing procedure in the GNSS receiver 1 according to the embodiment. The processing procedure (i.e., a series of steps from the start to the end) shown in the flow diagram of Fig. 2 is mainly a processing content executed by the positioning device 8 according to, for example, a program, and is repeatedly executed every time input information (observation information) is acquired at a predetermined time interval (specifically, for example, at one-second intervals).

[0018] The GNSS receiver 1 is a mechanism for acquiring position information by using a global positioning satellite system (GNSS: Global Navigation Satellite System). The GNSS receiver 1 has a GNSS antenna 2, a GNSS receiving unit 3, a data communication antenna 4, a data receiving unit 5, a satellite information acquiring unit 6, a reference station information acquiring unit 7, and a positioning device 8. The GNSS receiver 1 calculates and outputs its own position relative to other GNSS receivers.

[0019] The other GNSS receiver in the above is positioned as a “reference station” in the calculation of the relative position (i.e., relative positioning). The reference station is fixedly installed and its installation position is determined in advance (i.e., its installation position is known).

[0020] The GNSS antenna 2 receives signals transmitted from a number of satellites (e.g., GPS satellites). It receives radio waves (ranging signals; specifically, signals in which a PRN (pseudo-random noise) code and a navigation message are superimposed on a carrier wave of a specific frequency) and transfers them to the GNSS receiver 3.

[0021] Here, the multiple satellites that constitute the global positioning satellite system and transmit radio waves (ranging signals) received by the GNSS receiver 1 and the reference station are referred to as "multiple satellites Sx." When the multiple satellites Sx are to be distinguished from one another (in other words, when they are meant to be mutually different satellites), the subscript x is changed. Specifically, for example, "satellite Sj" and "satellite Sk" indicate that they are mutually different satellites.

[0022] The GNSS receiver 3 performs a predetermined process on the ranging signal transferred from the GNSS antenna 2 and outputs a signal in a format usable by the satellite information acquisition unit 6. For example, the GNSS receiver 3 converts the frequency of the ranging signal transferred from the GNSS antenna 2 to an intermediate frequency, performs an analog-to-digital conversion process to generate a digital signal, performs a demodulation process on the digital signal, and outputs the demodulated ranging signal (step S1).

[0023] The data communication antenna 4 receives data transmitted from the reference station and transfers it to a data receiving unit 5. Signals are transmitted and received between the GNSS receiver 1 and the reference station in accordance with a predetermined format.

[0024] The data transmitted from the reference station includes information on the ranging signals for each of the multiple satellites Sx received by the reference station via the GNSS antenna and information including the three-dimensional position of the reference station (referred to as "reference station satellite information"). The reference station satellite information is transmitted from the reference station at predetermined intervals and supplied to the GNSS receiver 1 (step S2).

[0025] The data receiving unit 5 performs a predetermined process on the data (that is, the reference station satellite information) transferred from the data communication antenna 4, and outputs a signal in a format that can be used by the reference station information acquiring unit .

[0026] The satellite information acquisition unit 6, the reference station information acquisition unit 7, and the positioning device 8 are configured within an arithmetic unit including, for example, a CPU (short for Central Processing Unit), memory, and input / output ports, and the CPU executes the necessary processing by referring to control programs and various data stored in the memory.

[0027] The satellite information acquisition unit 6 uses the ranging signal output from the GNSS receiver unit 3 at each predetermined period, and performs conventionally known arithmetic processing on the ranging signal as necessary to acquire information (referred to as "observation station observation data") regarding the satellite that transmitted the ranging signal for each of the multiple satellites Sx (step S3).

[0028] The satellite information acquisition unit 6 acquires, for each of the multiple satellites Sx, observation station observation data, such as the carrier phase (in other words, the carrier observation quantity) of the ranging signal received via the GNSS antenna 2, the pseudo-distance between the satellite that transmitted the ranging signal and the GNSS receiver 1 (in particular, the GNSS antenna 2), the reception time of the ranging signal, the signal strength of the ranging signal, and the three-dimensional position (x, y, z) of the satellite that transmitted the ranging signal.

[0029] The reference station information acquisition unit 7 uses information regarding the ranging signal received by the reference station, which is included in the reference station satellite information output from the data receiving unit 5, at each predetermined period, and performs conventionally known arithmetic processing on the ranging signal as necessary to acquire information regarding the satellite that transmitted the ranging signal for each of the multiple satellites Sx (referred to as "reference station observation data") (step S4).

[0030] The reference station information acquisition unit 7 acquires, for each of the multiple satellites Sx from which the reference station receives radio waves (ranging signals), reference station observation data such as the carrier phase (in other words, the carrier observation quantity) of the ranging signal received by the reference station, the pseudo-distance between the satellite that transmitted the ranging signal and the reference station (in particular, the GNSS antenna), the reception time of the ranging signal, the signal strength of the ranging signal, and the three-dimensional position of the reference station.

[0031] The observation data from the observation station and the reference station are synchronized for necessary items as necessary using GPS time or the like.

[0032] Note that some or all of the above-mentioned reference station observation data may be calculated in the reference station, and then included in the reference station satellite information and transmitted from the reference station to the GNSS receiver 1.

[0033] The positioning device 8 is a mechanism for calculating the relative position between the GNSS receiver 1 (in other words, the observation station) including the positioning device 8 and the reference station (i.e., the relative position of the GNSS receiver 1 with respect to the reference station) based on ranging signals transmitted from multiple satellites Sx, and has a real solution calculation unit 81, a first integer conversion unit 82, a second integer conversion unit 83, a fix solution position calculation unit 84, an integer solution variance adjustment unit 85, and a real solution update unit 86.

[0034] Specifically, the positioning device 8 according to this embodiment is a positioning device in which a reference station and the GNSS receiver 1 receive ranging signals transmitted from a plurality of satellites Sx, and calculate the relative position of the GNSS receiver 1 with respect to the reference station, and includes a real solution calculation unit 81 that calculates real-valued biases by solving a positioning equation in which the carrier phase of the ranging signal is used as an observable and the three-dimensional position of the GNSS receiver 1 and the real-valued bias of the carrier signal of the ranging signal are unknowns, a first integer conversion unit 82 that generates a first subset by using real-valued biases that satisfy a predetermined condition among the calculated real-valued biases, and converts the generated real-valued biases into integers in the processing in the first integer conversion unit 82. the real-valued bias calculated by the first integer conversion unit 82 is converted into an integer and the absolute value of the difference between the real-valued bias and the integer-valued bias converted into an integer by the first integer conversion unit 82 is smaller than a predetermined threshold value to generate a second subset by adding the real-valued bias that was not calculated as the first subset to the first subset, and converting the real-valued bias into an integer by the first integer conversion unit 82; a fix solution position calculation unit 84 that calculates the three-dimensional position of the GNSS receiver 1; an integer solution variance adjustment unit 85 that adjusts the variance of the integer value bias based on the success or failure of the integer conversion in the first integer conversion unit 82 and the success or failure of the integer conversion in the second integer conversion unit 83; and a real solution update unit 86 that updates the real-valued bias calculated by the real solution calculation unit 81 using the variance of the adjusted integer value bias.

[0035] The real solution calculation unit 81 calculates the three-dimensional position and real value bias of the GNSS receiver 1 by solving a positioning equation based on the observation information (specifically, the observation station observation data and the reference station observation data) (step S5). Here, the three-dimensional position of the GNSS receiver 1 obtained by the real solution calculation unit 81 is called a "float solution position (in other words, a position based on a real solution)" because it is calculated based on the real value bias.

[0036] First, the positioning equation used in the process of step S5 will be described.

[0037] First, the carrier phase (in other words, carrier observation quantity) φb_s [cycle] of the ranging signal when the radio wave from the satellite Ss is received at the reference station is expressed by the following formula 1. (Math. 1) λsφb_s = ρb_s+λsNb_s+c(dtb-dT_s)-Db_s+Tb_s Here, λs: wavelength of the carrier wave of satellite Ss [m] ρb_s: Geometric distance between the reference station and the satellite Ss [m] Nb_s: Real value bias [cycle] dtb: Clock error of the reference station [seconds] dT_s: Satellite clock error [seconds] Db_s: Ionospheric delay between the reference station and the satellite Ss [m] Tb_s: Tropospheric delay between the reference station and the satellite Ss [m] c: Speed ​​of light [m / sec] Subscript b: stands for reference station. The subscript s represents the satellite Ss.

[0038] Regarding the carrier phase of the ranging signal expressed according to Equation 1, when performing relative positioning using observation data from the observation station and observation data from the reference station, the double difference (double phase difference) φrb_jk of the carrier phase of the ranging signal for satellite Sj (reference satellite) and satellite Sk at the GNSS receiver 1 and the reference station is calculated as shown in Equation 2 below. (Math. 2) λsφrb_jk = ρrb_jk+λsNrb_jk Here, λs: wavelength of the carrier wave of satellite Ss ρrb_jk: double difference of geometric distances Nrb_jk: Double difference of real-valued bias The subscript r: represents the GNSS receiver 1. Subscript b: denotes reference station. The subscript j represents the satellite Sj (reference satellite). The subscript k represents the satellite Sk.

[0039] As can be seen from Equation 2, by calculating the double difference, the clock error of the receiving side (specifically, the GNSS receiver 1 and the reference station), the clock error of the satellite, the ionospheric delay, and the tropospheric delay superimposed on the observation data are offset.

[0040] In addition, when performing relative positioning, the double phase difference φrb_jk of the carrier phase (in other words, the carrier observable) of the ranging signal for satellite Sj (reference satellite) and satellite Sk between the GNSS receiver 1 and the reference station is calculated as shown in the following formula 3. (Math 3) φrb_jk = {φr_j-φb_j}-{φr_k-φb_k} Here, φr_j: Carrier phase of the ranging signal of satellite Sj (reference satellite) received by GNSS receiver 1 φb_j: Carrier phase of the ranging signal of satellite Sj (reference satellite) received at the reference station φr_k : Carrier phase of the ranging signal of satellite Sk received by GNSS receiver 1 φb_k: Carrier phase of the ranging signal of satellite Sk received at the reference station The subscript r represents the GNSS receiver 1. Subscript b: stands for reference station. The subscript j represents the satellite Sj (reference satellite). The subscript k represents the satellite Sk.

[0041] The double difference ρrb_jk of the geometric distance in Equation 2 is expressed as ρrb_jk = {ρr_j-ρb_j}-{ρr_k-ρb_k}, and the geometric distance ρb_j from the reference station to the satellite Sj (reference satellite) and the geometric distance ρb_k from the reference station to the satellite Sk are obtained from the three-dimensional positions of the satellites Sj and Sk calculated based on the satellite orbit information (in other words, ephemeris) and the known three-dimensional position of the reference station. Also, the geometric distance ρr_j from the GNSS receiver 1 to the satellite Sj (reference satellite) and the geometric distance ρr_k from the GNSS receiver 1 to the satellite Sk are obtained from the three-dimensional positions of the satellites Sj and Sk calculated based on the satellite orbit information and the approximate third-order position of the GNSS receiver 1.

[0042] In addition, the double difference Nrb_jk of the real-valued bias in Equation 2 is the real-valued bias Nr_j between the GNSS receiver 1 and the satellite Sj, the real-valued bias Nr_j between the reference station and the satellite Sj, Using b_j, the real-valued bias Nr_k between the GNSS receiver 1 and the satellite Sk, and the real-valued bias Nb_k between the reference station and the satellite Sk, Nrb_jk = {Nr_j-Nb_j}-{Nr_k-Nb_k}. The real-valued biases Nr_j, Nb_j, Nr_k, and Nb_k are referred to as "real-valued biases related to the double difference Nrb_jk."

[0043] Then, the double phase difference φrb_jk obtained according to Equation 3 is taken as the observable, and the three-dimensional position (precise three-dimensional position) of the GNSS receiver 1 and the real-valued bias related to the double difference Nrb_jk are taken as unknowns. The positioning equation (i.e., the positioning equation used in the processing of step S5) obtained by expanding the double difference ρrb_jk of the geometric distance in Equation 2 into a Taylor series expansion around the approximate third-order position of the GNSS receiver 1 is expressed as shown in Equation 4 below. (Equation 4) y = H x + ε

[0044] Equation 4 is a positioning equation constructed based on Equation 2, in which y is an observation vector having as its element the double phase difference φrb_jk obtained according to Equation 3, x is an unknown vector having as its elements the three-dimensional position (unknown) of the GNSS receiver 1 and a real-valued bias (unknown) related to the double difference Nrb_jk, and further, H is a design matrix that associates the observation vector y with the unknown vector x based on Equation 2. Also, ε is an observation noise term.

[0045] The real solution calculation unit 81 solves the positioning equation shown in Equation 4 to obtain an unknown vector x (specifically, the elements of which are the three-dimensional position of the GNSS receiver 1 and a real value bias related to the double difference Nrb_jk). The three-dimensional position of the GNSS receiver 1 obtained by the real solution calculation unit 81 is a float solution position because it is calculated based on the real value bias, and generally has a position accuracy of about several tens of centimeters.

[0046] The method for solving the positioning equation in the process of step S5 is not limited to a specific procedure or method, but may be, for example, a Kalman filter. Note that, in addition to the Kalman filter, for example, a sequential weighted least squares method such as the Gauss-Newton method may be applied.

[0047] The positioning process using the Kalman filter uses a group of equations shown in Equation 5 below.

number

[0048] The variables in formula 5 are as follows (note that "x^" has a ^ directly above the x). (This indicates that.) n: time x^(n|n-1): Estimated at time n using information up to time (n-1). Prior estimated state quantity x^(n|n) : Estimated at time n using information up to time n Posterior estimated state quantity y(n): Observed value F(n): State transition matrix M(n): Observation matrix K(n): Kalman gain P(n|n-1): prior error covariance matrix P(n|n): Posterior error covariance matrix Rε(n): Observation noise covariance matrix Rδ(n): system noise covariance matrix I: identity matrix T: stands for transposition.

[0049] Regarding the relationship between Formula 4 and Formula 5, the observation value y(n) in Formula 5 corresponds to the observation vector y in Formula 4, the observation matrix M(n) corresponds to the design matrix H in Formula 4, and the observation noise covariance matrix Rε(n) corresponds to the observation noise term ε in Formula 4.

[0050] According to Equation 5, the posterior estimated state quantity x^(n|n) is calculated by multiplying the observation prediction error by the Kalman gain K(n) and adding the result to the a priori estimated state quantity x^(n|n-1). Specifically, in the process of step S5, the three-dimensional position of the GNSS receiver 1 and the real-valued bias related to the double difference Nrb_jk are calculated according to Equation 5. The three-dimensional position of the GNSS receiver 1 calculated by the process of step S5 is based on the real-valued bias and is therefore a float solution position.

[0051] The real solution calculation unit 81 calculates and outputs the three-dimensional position of the GNSS receiver 1 and a real-valued bias related to the double difference Nrb_jk by, for example, applying a Kalman filter to solve the positioning equation (see Equation 4).

[0052] The first integer conversion unit 82 prepares a subset of the real-valued biases calculated by the process of step S5, and converts the real-valued biases of the subset into integers (step S6).

[0053] Here, the ranging signal transmitted from the satellite is affected by disturbances such as momentary interruptions, and the quality of the ranging signal is degraded, which in turn degrades the quality of the carrier phase. For this reason, in the process of step S6, a first subset of real-valued biases is generated by eliminating the real-valued biases affected by the degradation of the quality of the carrier phase, and the real-valued biases of this first subset are converted into integers. The first subset of real-valued biases is generated by, for example, selecting only those real-valued biases calculated in the process of step S5 that have a good signal strength (dB / Hz; specifically, the carrier power-to-noise ratio, which is the noise power per 1 Hz band) of the received ranging signal. Alternatively, the first subset may be composed of only real-valued biases calculated using observation station observation data on satellites with high elevation angles in the GNSS receiver 1. Furthermore, the selection conditions for the real-valued biases constituting the first subset are not limited to specific conditions, and appropriate conditions may be appropriately set, taking into consideration that a certain level of quality of the ranging signal can be expected to be maintained.

[0054] The first integer conversion unit 82 converts the first subset into an integer number, for example, The integer bias (bias of the integer solution; i.e., wave number) with the smallest error is obtained as a first candidate for the real-value bias (bias of the real solution) that constitutes the set, and the integer bias with the next smallest error is obtained as a second candidate. Although the method for identifying an integer solution with a small error is not limited to a specific procedure or method, for example, the integer least squares method is used, and in particular, the LAMBDA method, which aims to decorrelate the real-value bias group and narrows the search space of the integer solution to identify a solution (in other words, the integer bias group for the real-value bias group).

[0055] The first integer conversion unit 82 then determines the reliability of the integer conversion of the integer value bias group (i.e., the integer converted real value bias group) obtained by the processing of step S6, and if a reliable integer solution is obtained, it determines that the integer conversion of the real value bias corresponding to the integer solution has been successful (step S7: Yes).

[0056] The method for determining the reliability of the integer solution is not limited to a specific procedure or method, but may be, for example, a ratio test. The ratio, which is an index used in the ratio test, is the ratio of the distance between the real value bias and the second candidate of the integer value bias to the distance (called the "norm") between the real value bias and the first candidate of the integer value bias (in other words, the real value bias converted into an integer). Generally, the higher the ratio, the higher the reliability of the first candidate of the integer value bias. Therefore, it is conceivable to set an appropriate threshold value, and when the ratio is greater than the threshold value, to determine that the reliability of the integer value bias is high and to determine that the integer conversion of the real value bias corresponding to the integer value bias has been successful. Here, when it is determined that the integer conversion of the real value bias has been successful, the first candidate of the integer value bias is adopted as the real value bias converted into an integer.

[0057] Then, if it is determined that the integer conversion of the real-valued bias group of the first subset is not successful (step S7: No), the first integer conversion unit 82 advances the positioning processing procedure to the end of the flow chart of Figure 2, outputs the three-dimensional position of the float solution obtained by the real solution calculation unit 81 from the positioning device 8, and when a new observation is input, restarts the positioning processing procedure from the start of the flow chart of Figure 2.

[0058] On the other hand, if it is determined that the integer conversion of the real-valued biases of the first subset has been successful (step S7: Yes), the first integer conversion unit 82 then determines whether all the real-valued biases have been converted to integers (step S8).

[0059] Here, all real-valued biases refer to all real-valued biases calculated in the calculation of real-valued biases (step S5), and it is determined whether all real-valued biases calculated in the process of step S5 have been converted to integers in the process of step S6. That is, in the process of S8, it is determined whether there are any real-valued biases among the real-valued biases calculated in the process of step S5 that were not selected as the first subset in the process of step S6.

[0060] Then, if all the real-valued bias calculated in the processing of step S5 have been converted to integers (step S8: Yes), the first integer conversion unit 82 outputs the set of integer-valued bias obtained in the processing of step S6 (in other words, the set of integer-converted real-valued bias) to the fix solution position calculation unit 84, and proceeds with the positioning processing procedure to processing of step S10.

[0061] On the other hand, when some of the real-valued bias calculated in the process of step S5 have not been converted to integers (step S8: No), the first integer conversion unit 82 advances the positioning process procedure to the process of step S9.

[0062] The second integer conversion unit 83 generates a new subset of real-valued biases (called the "second subset" to distinguish it from the first subset) based on the result of the processing in step S6, and converts the real-valued biases of the second subset into integers (step S9). Figure 3 is a flow diagram showing the processing procedure in the second integer conversion unit 83, and is a flow diagram showing the details of the processing in step S9 in the flow diagram of Figure 2.

[0063] The second integer conversion unit 83 first separates the real-valued biases obtained by the real-valued solution calculation unit 81 (step S5) into biases that were subject to integer conversion processing in the processing of step S6 (in other words, biases that belong to the first subset) and biases that were not subject to integer conversion processing (in other words, biases that do not belong to the first subset), and calculates the bias error of the biases that do not belong to the first subset relative to the biases that belong to the first subset (step S9-1).

[0064] The bias error Nerr is calculated from the inter-receiver single difference Nrb_j of the bias for a satellite j belonging to the first subset and the inter-receiver single difference Nrb_k of the bias for a satellite k not belonging to the first subset according to the following formulas 6-1 and 6-2. (Number 6-1) Nrb_jk = Nrb_j-Nrb_k (Number 6-2) Nerr = |Nrb_jk-round(Nrb_jk)| Here, Nrb_j: The single difference in bias between receivers for satellite j in the first subset [cycle] Nrb_k: Single difference between receivers in bias for satellite k not belonging to the first subset [cycle] Nrb_jk: Bias double difference [cycle] Nerr: Bias error [cycle] round(·): A function that rounds the argument to the nearest integer. ||: Absolute value

[0065] The bias double difference Nrb_jk calculated by Equation 6-1 is the error of the inter-receiver single difference Nrb_k (=Nr_k-Nb_k) of bias for satellite k not belonging to the first subset, based on the inter-receiver single difference Nrb_j (=Nr_j-Nb_j) of bias for satellite j belonging to the first subset that was successfully converted to an integer (in other words, an index value for evaluating the quality of the inter-receiver single difference Nrb_k). The bias error Nerr calculated by Equation 6-2 is the decimal part of the bias double difference Nrb_jk calculated by Equation 6-1, and the bias error Nerr ranges from 0.0 to 0.5 [cycle]. If the bias error Nerr is 0.0 [cycle], the error in the inter-receiver single difference Nrb_k of the bias for a satellite k that does not belong to the first subset can be determined to be small, and if the bias error Nerr is close to 0.5 [cycle], the error in the inter-receiver single difference Nrb_k of the bias for a satellite k that does not belong to the first subset can be determined to be large.

[0066] Here, in the process of step S6, the bias of the first subset converted to integers is obtained in the form of double differences Nrb_jk. Then, when the biases Nrb_j, Nrb_k of the single differences between receivers are calculated from the integer biases of the double differences, they become real values ​​rather than integer values. That is, the bias of the first subset is an integer bias in the case of double differences, and a real bias in the case of single differences. In the explanation related to Equation 6-1 and Equation 6-2, the notation "bias" is also used simply in consideration of such characteristics.

[0067] The second integer conversion unit 83 calculates the bias error Nerr of the biases not belonging to the first subset using Equation 6-1 and Equation 6-2, and then, if the bias error Nerr is smaller than a predetermined bias error threshold, generates a bias group of a second subset by adding the biases not belonging to the first subset and having the smaller bias error Nerr to the first subset (step S9-2).

[0068] That is, the biases in the second subset are a collection of biases that is a combination of the set of biases in the first subset and the set of biases that do not belong to the first subset and that are determined to have a small bias error Nerr.

[0069] The bias error threshold is not limited to a specific value, and is appropriately set to an appropriate value, taking into consideration the magnitude of the error that is considered not to significantly impede the success of integerization when added to a set of real-valued biases that have been successfully integerized. The bias error threshold may be set to any value within the range of about 0.05 to 0.25 [cycle], for example, and may be set to about 0.15 [cycle] in particular.

[0070] The second integer conversion unit 83 then converts the set of real-valued biases of the second subset into integers (in other words, obtains a set of integer-valued biases for the set of real-valued biases of the second subset) (step S9-3).

[0071] The second integerization unit 83, for example, determines an integer bias having the smallest error for the real biases constituting the second subset generated by the process of step S9-2 as a first candidate, and determines an integer bias having the second smallest error as a second candidate. Although the method for identifying an integer solution having a small error is not limited to a specific procedure or method, for example, the integer least squares method is used, and in particular, the LAMBDA method, which aims to decorrelate the real biases and narrows the search space of integer solutions to identify a solution (in other words, an integer bias group for the real bias group).

[0072] The second integer conversion unit 83 then determines the reliability of the integer conversion of the integer value bias group (i.e., the integer converted real value bias group) obtained by the processing of step S9-3, and if a reliable integer solution is obtained, determines that the integer conversion of the real value bias corresponding to the integer solution has been successful (step S9-4).

[0073] Although the method for determining the reliability of the integer solution is not limited to a specific procedure or method, for example, a ratio test may be used, as in the process of step S7. Then, when it is determined that the reliability of the integer value bias is high based on the ratio, it may be determined that the real value bias corresponding to the integer value bias has been successfully converted into an integer.

[0074] Then, if it is determined that the integer conversion of the real-valued bias group of the second subset has been successful (step S9-4: Yes), the second integer conversion unit 83 outputs a set of integer-valued biases obtained by integer converting the real-valued bias group of the second subset (in other words, a set of integer-converted real-valued biases) to the fix solution position calculation unit 84 (step S9-5).

[0075] On the other hand, if it is not determined that the integer conversion of the real-valued bias group of the second subset has been successful (step S9-4: No), the second integer conversion unit 83 outputs to the fix solution position calculation unit 84 a set of integer-valued biases obtained by integer converting the real-valued bias group of the first subset that was the target of the processing in step S6 (step S9-6).

[0076] The fix solution position calculation unit 84 corrects the float solution position calculated in the processing of step S5 using a set of integer biases output from the first integer conversion unit 82 or the second integer conversion unit 83, and calculates the three-dimensional position of the GNSS receiver 1 (step S10). The three-dimensional position of the GNSS receiver 1 calculated in the processing of step S10 is calculated based on the integer biases and is therefore called a "fix solution position (in other words, a position based on an integer solution)", and generally has a position accuracy of about several centimeters.

[0077] In response to the results of the processes in steps S8 and S9-4, the integer solution variance adjustment unit 85 adjusts the variance value of the integer value bias used in the real solution update unit 86 at the subsequent stage (step S11).

[0078] Specifically, for example, when it is determined that the environment related to satellite radio wave reception is unstable, the integer solution dispersion adjustment unit 85 adjusts to increase the dispersion value of the integer value bias used by the subsequent real number solution update unit 86.

[0079] a) When Yes in the process of Step 8 It is determined that there is no real number value bias that cannot be integerized and it is the most stable situation, and the dispersion value of the integer value bias is set to a1 [cycle 2 . b) When Yes in Step S9-4 (when executing the process of Step S9-5) Since the integerization of the newly added real number value bias in the process of Step S9-3 was successful, it is determined that although it is an unstable situation where momentary disconnection or cycle slip is likely to occur, it is a more stable situation than the following c), and the dispersion value of the integer value bias is set to a2 [cycle 2 . c) When No in Step S9-4 (when executing the process of Step S9-6) Since the integerization of the newly added real number value bias in the process of Step S9-3 failed, it is a situation where momentary disconnection or cycle slip is likely to occur, and it is determined that it is a more unstable situation than the above b), and the dispersion value of the integer value bias is set to a3 [cycle 2 .

[0080] The coefficients a1, a2, and a3 in the above are the dispersion values of the integer value bias. Since the dispersion value becomes a small value in a stable situation, they are set as real numbers satisfying the relationship 0.0 < a1 < a2 < a3, and none of them are limited to specific values. For example, the coefficient a1 can be set to any value in the range of about 0.0001 to 0.01, and particularly can be set to about 0.001. Also, the coefficient a2 can be set to any value in the range of about 0.02 to 0.7, and particularly can be set to about 0.1. Further, the coefficient a3 can be set to any value in the range of about 0.8 to 1.0, and particularly can be set to about 1.0.

[0081] The real solution update unit 86 obtains an unknown vector x having elements that are real-valued biases related to the three-dimensional position of the GNSS receiver 1 and the double difference Nrb_jk by solving the positioning equation shown in Equation 7 below. (Number 7) y B = H B x+ε B

[0082] In the formula 7, y B is an observation vector having as its elements the difference between the integer bias calculated by the first integer conversion unit 82 or the second integer conversion unit 83 and the real bias before integer conversion, x is an unknown vector having the same configuration as the unknown vector x in Equation 4 and having as its elements the three-dimensional position (unknown) of the GNSS receiver 1 and the real bias (unknown) related to the double difference Nrb_jk, and H B is the observation vector y B is the design matrix that relates the unknown vector x to the unknown vector x. Also, ε B is an observation noise term, and the variance of the integer bias calculated by the integer solution variance adjustment unit 85 is applied to the variance of the observation noise term.

[0083] The real solution update unit 86 obtains the unknown vector x by solving the positioning equation of Equation 7 using the Kalman filter equation shown in Equation 5. That is, the real solution update unit 86 solves the positioning equation of Equation 7 based on the integer bias obtained by the first integer conversion unit 82 or the second integer conversion unit 83 and the variance of the integer bias obtained by the integer solution variance adjustment unit 85, thereby updating the real solution biases related to the three-dimensional position of the GNSS receiver 1 and the double difference Nrb_jk, which are elements of the unknown vector x (step S12).

[0084] In the process of step S12, when solving the positioning equation of Equation 7 using the Kalman filter equation shown in Equation 5, the variance value of the integer bias element constituting the observation noise covariance matrix Rε(n) of Equation 5 is applied with the variance value of the integer bias adjusted in the process of step S11.

[0085] The method for solving the positioning equation in the process of step S12 is not limited to a specific procedure or method, but may be, for example, a Kalman filter. Note that, other than the Kalman filter, for example, a sequential weighted least squares method such as the Gauss-Newton method may be applied.

[0086] The real-valued bias updated in step S12 is used as feedback when a new observation is input and the process is restarted from step S1. Specifically, after the current process is completed and the time is updated and the process is restarted from step S1, the real-valued bias updated in step S12 at the previous timing (i.e., time n-1) is used for the pre-estimated state quantity x^(n-1|n-1) in formula 5 in step S5. In other words, by using the information on the real-valued bias updated in step S12 as feedback, the calculation result of the real-valued bias at the next timing becomes more accurate. This feedback control makes the three-dimensional position and double difference Nrb_jk of the GNSS receiver 1, which are calculated as unknowns, more accurate (in other words, the accuracy is improved successively).

[0087] The positioning device 8 then outputs the determined float solution position or fixed solution position of the GNSS receiver 1 to the outside, and when new observations are input, restarts the positioning processing procedure from the start point of the flow chart of FIG.

[0088] According to the positioning device 8 as described above, for the real-valued biases obtained in the processing of step S5 that were not subject to the integer conversion processing of step S6, a bias error Nerr is calculated in the processing of step S9-1, and a second subset is generated by adding the real-valued biases whose bias error Nerr satisfies a predetermined condition to the first subset, and the real-valued biases of the second subset are converted to integers (step S9-3). Therefore, the number of integer-valued biases can be increased, and by performing positioning using a large number of integer-valued biases, it is possible to obtain a highly accurate positioning position.

[0089] Specifically, as shown in FIG. 4, assuming that an instantaneous interruption occurs at time A, the bias error increases rapidly due to the influence of cycle slip caused by the instantaneous interruption, and then gradually decreases as a general trend, repeating small increases and decreases, and converging at time C. Conventionally, the real-value bias is not converted to an integer until time C is reached, since the bias error of the real-value bias is converged before being converted to an integer. In contrast, according to the positioning device 8 as described above, even if the real-value bias is affected by a deterioration in the quality of the carrier wave due to an instantaneous interruption, the bias error of the real-value bias is calculated in the process of step S9-1, so that time B at which the bias error of the real-value bias instantaneously decreases to the convergence level is detected, and the real-value bias at time B is converted to an integer. This makes it possible to increase the number of integer-value biases, and to perform positioning using many integer-value biases, thereby making it possible to obtain a highly accurate positioning position.

[0090] According to the positioning device 8 as described above, the variance of the integer bias is adjusted based on the integerization situation (step S11), and the real value bias is updated using the adjusted variance of the integer bias (step S12). Therefore, in an unstable situation where a momentary interruption or cycle slip is likely to occur, the variance of the integer bias is adjusted to be large, so that the real value bias can be updated with an appropriate variance of the integer bias. This reduces the effect of the integer bias with a large variance on the update of the real value bias, and therefore the update error of the real value bias can be suppressed even if the surrounding environment changes significantly, and a stable and highly accurate positioning position can be continuously obtained. It becomes possible to obtain the above.

[0091] Although the embodiment of the present invention has been described above, the specific configuration is not limited to the above embodiment, and even if there are design changes and the like within the scope of the gist of the present invention, they are included in the present invention.

[0092] For example, in the above embodiment, the variance of the integer value bias is adjusted based on the integer conversion situation (step S11), and the adjusted variance of the integer value bias is used to update the real value bias (step S12). However, the processes of steps S11 and S12 are not essential configurations for this invention, and it is also possible to perform the process of step S10 after the process of step S8 or the process of step S9, and then return to the process of step S1.

[0093] Furthermore, the main feature of this invention is the processing of step S9 by the second integer conversion unit 83, and a subordinate feature is the processing of step S11 by the integer solution variance adjustment unit 85. The specific configurations of the GNSS receiver 1 and the positioning device 8 are not limited to the example shown in Figure 1, and the processing contents other than the above-mentioned features are not limited to the processing contents in the above-mentioned embodiment. [Industrial Applicability]

[0094] This invention can be applied, for example, to the fields of in-vehicle navigation systems, vehicle control and automatic driving systems. In addition, this invention is applicable not only to the RTK (short for Real Time Kinematic) positioning method using reference station information, but also to a Precise Point Positioning (PPP) method that uses the centimeter-level augmentation service (CLAS) broadcast from Japan's quasi-zenith satellites instead of reference station information to suppress errors in observation data from observation stations. It can also be applied to the GPS-RTK (Real-time Kilometre-Kind) positioning method and the High Precision Point Precise Positioning (PPP) method, which uses precise satellite almanac (satellite position, satellite clock) and tropospheric delay information provided by augmentation information providers instead of reference station information to suppress errors in observation data from observation stations. [Explanation of symbols]

[0095] 1 GNSS receiver 2 GNSS antennas 3 GNSS receiver 4 Data communication antenna 5 Data receiving section 6 Satellite information acquisition section 7 Reference station information acquisition section 8 Positioning Device 81 Real solution calculation section 82 First integerizer 83 Second integerizer 84 fix solution position calculation part 85 Integer solution variance adjustment unit 86 Real solution update section

Claims

1. A positioning device in which a reference station and a receiver receive ranging signals transmitted from a plurality of satellites and calculate a relative position of the receiver with respect to the reference station, a real solution calculation unit that calculates the real-valued bias by solving a positioning equation having a carrier phase of the ranging signal as an observable and a three-dimensional position of the receiver and a real-valued bias of the carrier signal of the ranging signal as unknowns; a first integer conversion unit that converts the real-valued biases calculated by the first integer conversion unit into integers by generating a first subset based on the real-valued biases that satisfy a predetermined condition; a second integer conversion unit that converts the real-valued biases calculated by the first integer conversion unit into integers by adding to a first subset the real-valued biases that are not subject to integer conversion in the first integer conversion unit and have a fractional part whose difference from the real-valued biases that are subject to integer conversion in the first integer conversion unit has an absolute value smaller than a predetermined threshold value, If the first subset includes all of the real-valued biases, or if the integerization of the second subset is not successful, correct the three-dimensional position of the receiver calculated in the real solution calculation unit using a set of integer-valued biases output from the first integerization unit, and calculate the relative position of the receiver with respect to the reference station; If the first subset does not include all of the real-valued biases and the integerization of the second subset is successful, a set of integer-valued biases output from the second integerization unit is used to correct the three-dimensional position of the receiver calculated in the real solution calculation unit, and calculate the relative position of the receiver with respect to the reference station. A positioning device characterized by:

2. an integer solution variance adjustment unit that adjusts the variance of the integer value bias based on a success or failure status of the integer conversion in the first integer conversion unit and a success or failure status of the integer conversion in the second integer conversion unit; and a real solution update unit that updates the real value bias calculated by the real solution calculation unit using the adjusted variance of the integer value bias.

2. The positioning device according to claim 1 .

3. Solving the positioning equation by applying a Kalman filter; 3. The positioning device according to claim 1 or 2.

Citation Information

Patent Citations

  • Carrier phase relative determination device and determination system

    JP2001013237A

  • Satellite positioning / computing method and satellite navigating / positioning device using the same

    JP2003270319A

  • Carrier phase type mobile positioning system

    JP2009025049A

  • Inter-moving object interference positioning apparatus and method for moving object

    JP2009270927A

  • Positioning method and positioning system

    JP2016057239A