KR101941132B1 - Apparatus and method for extending available area of regional ionosphere map - Google Patents

Abstract

The present invention relates to a method of expanding a usable area of a local ionospheric map, and a method of expanding a usable area of a local ionospheric map includes: (a) inputting input data including a plurality of past time information and a past ionospheric delay value associated with the plurality of past time information; Determining a target data for an extension point of the unavailable region at an expansion time point to generate a model; And (b) performing learning using the model, and calculating an extrapolated expanded ionospheric delay value with respect to the extension point at the expansion time.

Description

FIELD OF THE INVENTION The present invention relates to an apparatus and method for expanding available area of ionospheric map,

The present invention relates to an apparatus and method for expanding a usable area of a local ionospheric map.

In case of using a single frequency receiver, the signal transmitted from the satellite passes through the ionosphere and a signal delay occurs. Errors due to the ionosphere distributed over 70 ~ 1000 km in the sky will increase the user 's estimation error.

As a method of generating correction information for the ionospheric delay, there is a method of modeling the ionospheric delay value as a function according to the user's position based on the observation value, a method of setting a lattice point and a lattice point generating a vertical ionospheric delay value of each lattice point There is a point method. The function method is a method of generating the coefficient of the modeled function by expressing the ionospheric delay value estimated using the dual frequency at the observatory as a function of time, latitude and longitude, and by using the least squares method, There is a disadvantage in that there is a difference from the actual value. On the other hand, the lattice point method is a method of estimating the ionospheric delay value at the lattice point by interpolation with the estimated ionospheric delay value at each observatory. Obtaining a similar value as the actual value However, there is a disadvantage in that the accuracy is lowered toward the edge of the lattice point.

As a method of correcting the ionospheric error, ionospheric correction information transmitted in the navigation message can be used. The Klobuchar model transmitted from GNSS satellites is a function method and can be used to correct ionospheric error using the ionospheric pierce point (IPP) of the user receiver and the coefficients of the transmitted Klobuchar model. This method has the advantage of correcting the ionospheric error in real time and correcting the ionospheric error anywhere, irrespective of the user's location. However, compared to the actual ionospheric delay, the root mean square (RMS) error is about 2 m , There is a disadvantage that the correction effect is insufficient.

Another ionospheric delay correction method is to use differential GNSS (Differential GNSS: DGNSS). The above method is a lattice point method in which a user can numerically interpolate an ionospheric delay value provided in a lattice point form and use it for correction. In this method, the ionospheric error can be corrected to within 50 cm. However, in order to use the differential satellite navigation system, there is a restriction that the user must exist within about 100 km from the reference station. This is because, when the distance between the user and the reference station becomes long, the correlation between the correction information generated at the reference station and the error generated at the user's position becomes poor.

Another ionospheric delay correction method is to calibrate using the global ionospheric map provided by IGS (International GNSS service). Global ionospheric maps provide ionospheric delay values for global grid points using data collected from global observatories and provide ionospheric delay values at 2.5 ° latitude and 5 ° longitude. IGS global ionospheric maps generate ionospheric delay correction information across the globe, so they have the advantage of correcting ionospheric errors with accuracy within the range of 0.3 to 1.2 m at any location, It takes about 10 ~ 12 days to combine the ionospheric map and it can not be used in real time.

Another method is to create a local ionospheric map and extrapolate it to extend the local ionospheric map. The local ionospheric map is a method of generating an ionospheric map in the form of a lattice point around an existing observatory and acquiring an ionospheric delay value by extrapolating it to a user outside the local ionospheric map. For extrapolations, for example, there is a method of expanding the usable area of the ionospheric map using the biharmonic spline extrapolation method. In this method, the accuracy of an average of 30 cm error in extrapolation of unavailable ionospheric delay values at 5 ° latitude / longitude from the ionospheric map availability area at active solar activity was shown. However, there is a disadvantage that the extrapolation error increases up to 1 m in spring and autumn when the ionospheric delay value increases.

Another extrapolation method is to use an artificial neural network. This uses the solar and geomagnetic activity index, time information, and extrapolated location information measured at the ground station as the inputs to the artificial neural network. In this method, the non-available area 5 ° away from the nearest ground station at latitude / longitude The accuracy of extrapolation of ionospheric delay values was 40 cm. This was similar to the accuracy of the biharmonic spline extrapolation method described above.

Meanwhile, since the extrapolation methods of the conventional ionospheric map extrapolation methods are performed using the solar activity and the geomagnetic activity information of the extrapolated point (present time) at which the extrapolation is performed, the solar and geomagnetic activity indexes are not measured There is a problem that extrapolation can not be performed at the present time point. That is, in the past, extrapolation could only be performed as post-processing after measurement of the sun and geomagnetic activity index at the extrapolation time point.

On the other hand, the background technology of the present application is disclosed in Korean Patent Registration No. 10-1607082. The Korean Patent Publication discloses only extension of the service area of differential satellite navigation, which has a limited range of extrapolation and can not be used to expand the ionospheric usable region in the world ionospheric map. In addition, since the Korean Registered Patent Bulletin performs extrapolation in the form of a spatial extrapolation using only the ionospheric delay value generated in the reference station and extrapolates the linear extrapolation, a large number of variable elements affecting the ionospheric delay There is a disadvantage in that the extrapolation performance is poor.

There have also been some studies related to the extension of the local ionospheric map. For example, ["Expansion of Ionospheric Compensation Information Area Using Extrapolation Technique", Journal of Korean Aircraft Operation, pp. 74-81, 2014. Kim, Jung-Rae, Kim, Min-gyu] This paper describes a study on local ionospheric map expansion using extrapolation techniques such as Kriging and Biharmonic spline. The above paper uses only the ionospheric delay value in the ionospheric map area as the input used for extrapolation and does not take into consideration a number of variable factors affecting the ionospheric delay, resulting in poor extrapolation performance.

In addition, a study on the expansion of the local ionospheric map using artificial neural networks has been carried out [[Extend ionospheric correction coverage area by using a neural network method], International Journal of Aeronautical and Space Science, pp. 64-72, 2016, Kim, Mingyu, Kim Jeongrae] and ["Regional GPS TEC Modeling: Attempted spatial and temporal extrapolation of TEC using neural networks", Journal of Geophysical Research, 2011, pp. 1-14, John Bosco Habarulema, Lee-Anne McKinnell, and Ben D. L. Opperman. The two papers above are extrapolation structures after the measurement of the sun and geomagnetic activity indexes are performed, and extrapolation can not be performed unless measurements of sun and geomagnetic activity indexes are performed. Therefore, the above two papers have a disadvantage in that it is impossible to extrapolate in real time the change of the ionospheric delay value in real time.

The present invention has been made to overcome the above-described problems of the conventional art, and it is an object of the present invention to overcome the temporal and spatial limitations of the ionospheric error application method using the existing lattice point method and to use the lattice point type ionospheric error correction information It is an object of the present invention to provide an apparatus and method for expanding the available region of ionospheric map which provides accurate correction information to the ionospheric delay value to a user who can not.

The present invention provides an apparatus and method for expanding a usable area of a local ionospheric map that provides improved extrapolation performance by extrapolating a plurality of variable factors affecting ionospheric delay The purpose is to do.

SUMMARY OF THE INVENTION It is an object of the present invention to provide an apparatus and method for expanding a usable area of a local ionospheric map capable of real-time extrapolation without measuring sun and geomagnetic activity indexes at the present time.

It is to be understood, however, that the technical scope of the embodiments of the present invention is not limited to the above-described technical problems, and other technical problems may exist.

According to an aspect of the present invention, there is provided a method for expanding an available area of a local ionospheric map according to an embodiment of the present invention, including: (a) obtaining a plurality of past time information and a past ionospheric delay Determining a target data for an extension point of the unavailable area at an expansion time and generating a model; And (b) performing learning using the model, and calculating an extrapolated expanded ionospheric delay value with respect to the extension point at the expansion time.

In addition, in the step (a), the input data may further include past sun / geomagnetic activity information associated with the plurality of past time information, and the step (b) It is possible to perform extrapolation using solar / geomagnetic activity information of the most recent point in the past solar / geomagnetic activity information associated with the plurality of past time information as solar / geomagnetic activity information at the expansion point .

Also, the extension time may be the current time, and the extrapolation in step (b) may be real-time extrapolation.

Also, in the step (a), the past ionospheric delay value may include past ionospheric delay values for the available region and past ionospheric delay values for unavailable regions.

The past ionospheric delay value for the unavailable region may be a postprocess ionosphere delay value or a past ionospheric delay value calculated through post-processing based on a past observation value at a past time point corresponding to any one of the plurality of past time information, And may include the extended ionospheric delay value calculated through the step (b) with the time point being an extension time point.

Also, the post-treatment ionospheric delay value may include an ionospheric delay value obtained from the global ionospheric map provided in non-real-time through post-processing in IGS.

In addition, in the step (a), the plurality of past time information corresponds to a predetermined past period based on the expansion time, and the first time period after the first execution of the steps (a) and (b) Wherein when the second execution is performed after a predetermined time elapses from the execution time point and after the elapse of the predetermined time, the input data of step (a) The past time information may be updated to a plurality of past time information corresponding to a preset past period on the basis of the past time information.

In the step (a), the model is a model including input data for a SVM (Support Vector Machine) and target data. In the step (b), the learning is performed based on a SVM .

Meanwhile, the available area expansion apparatus of the local ionospheric map according to an embodiment of the present invention determines input data including a plurality of past time information and a past ionospheric delay value associated with the plurality of past time information, A model generation unit for generating a model by determining target data for an extension point of a non-available area; And a calculator performing learning using the model and calculating an extrapolated expanded ionospheric delay value with respect to the expansion point at the expansion time.

The input data in the model generating unit may further include past sun / geomagnetic activity information related to the plurality of past time information, and the calculating unit may acquire solar / geomagnetic activity information at the expansion time Extrapolation can be performed using solar / geomagnetic activity information of the most recent point among the past solar / geomagnetic activity information associated with the plurality of past time information as solar / geomagnetic activity information at the expansion point.

Also, the extension time may be the current time, and the extrapolation performed by the calculation unit may be real-time extrapolation.

Also, the past ionospheric delay value in the model generator may include past ionospheric delay values for the available region and past ionospheric delay values for unavailable regions.

The past ionospheric delay value for the unavailable region may be a postprocess ionosphere delay value or a past ionospheric delay value calculated through post-processing based on a past observation value at a past time point corresponding to any one of the plurality of past time information, And may include an extended ionospheric delay value calculated through the calculation unit with the time point as an expansion time point.

Also, the post-treatment ionospheric delay value may include an ionospheric delay value obtained from the global ionospheric map provided in non-real-time through post-processing in IGS.

It is preferable that the plurality of past time information in the model generation unit corresponds to a predetermined past period based on the extension time point, and the plurality of past time information corresponds to a predetermined period of time after the first execution by the model generation unit and the calculation unit. When the second execution is performed after the predetermined time elapses from the time point and after the predetermined time elapsed after the elapse of the predetermined time, The past time information may be updated to a plurality of past time information corresponding to a past predetermined period.

Also, in the model generation unit, the model is a model including input data and target data for SVM (Support Vector Machine), and in the calculation unit, the learning may be performed based on SVM (Support Vector Machine) .

The above-described task solution is merely exemplary and should not be construed as limiting the present disclosure. In addition to the exemplary embodiments described above, there may be additional embodiments in the drawings and the detailed description of the invention.

According to the above-mentioned problem solving means of the present invention, extrapolation is performed in consideration of a plurality of variable factors (for example, time information, ionospheric delay value for available and unavailable regions, solar / geomagnetic activity information) , It is possible to provide a more improved extrapolation performance.

According to the above-mentioned problem solving means of the present invention, when the solar / geomagnetic activity information at the expansion time can not be obtained, the solar / geomagnetic activity information of the most recent point among the past solar / May be used to perform extrapolation by performing extrapolation using solar / geomagnetic activity information at the extension time.

According to the above-mentioned problem solving means of the present invention, it is possible to learn the ionospheric delay value of the unavailable region according to the change of the input data by using a plurality of variable elements, and to learn the ionospheric delay value at the time when the ionospheric delay value of the non- Extrapolation can be performed on the basis of the information, which can be used to increase the extent of extrapolation and accurately reflect changes in the ionosphere.

We are unable to install reference stations such as marine areas on a local ionospheric map of lattice points generated using only the measurement information of a regular observation station, or the ionosphere correction information can be provided because a reference station is not installed due to various factors It is possible to provide ionospheric correction information in real time in an area where there is no ionospheric correction.

Conventionally, the ionospheric correction information provided by the function method was extrapolated only to the post-processing due to the presence of input that is difficult to measure in real time, such as accuracy or inaccuracy of the solar / geomagnetic activity, By extrapolating the past data before the extrapolation is performed, it is possible to estimate the ionospheric delay value of the unavailable region by extrapolating in real time.

 Since we can estimate the ionospheric delay value by expanding the local ionospheric map without needing to install additional observation stations for estimating the ionospheric delay value for the low economic or difficult to install regions or countries, The cost for additional installation can be reduced.

We can generate ionospheric calibration information for an area where there is no observatories in all areas using the satellite navigation system. For example, the present invention can extend the ionospheric availability area of a differential satellite navigation system or extend the area of local ionospheric maps generated at domestic stations.

However, the effects obtainable here are not limited to the effects as described above, and other effects may exist.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a diagram schematically illustrating a configuration of an available area expansion device of a local ionospheric map according to an embodiment of the present invention; FIG.
FIG. 2 shows an example of setting a usable area and a non-usable area for performance evaluation of a usable area expanding device of a local ionospheric map according to an embodiment of the present invention.
FIG. 3 is a diagram showing an extrapolated expanded ionospheric delay value by the available area expansion device of the local ionospheric map according to an embodiment of the present invention and an ionospheric delay value extracted from the IGS global ionospheric map.
4 is a graph showing the difference between the two ionospheric delay values in FIG.
5 is a schematic flowchart illustrating a method of expanding a usable area of a local ionospheric map according to an embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings so that those skilled in the art can easily carry out the present invention. It should be understood, however, that the present invention may be embodied in many different forms and should not be construed as limited to the embodiments set forth herein. In the drawings, the same reference numbers are used throughout the specification to refer to the same or like parts.

Throughout this specification, when an element is referred to as being "connected" to another element, it is intended to be understood that it is not only "directly connected" but also "electrically connected" or "indirectly connected" "Is included.

It will be appreciated that throughout the specification it will be understood that when a member is located on another member "top", "top", "under", "bottom" But also the case where there is another member between the two members as well as the case where they are in contact with each other.

Throughout this specification, when an element is referred to as "including " an element, it is understood that the element may include other elements as well, without departing from the other elements unless specifically stated otherwise.

Hereinafter, the available area expansion device (hereinafter referred to as 'the expansion device') of the local ionospheric map according to an embodiment of the present invention will be described in detail.

FIG. 1 is a diagram schematically showing a configuration of an available area expansion apparatus 100 of a local ionospheric map according to an embodiment of the present invention.

Referring to FIG. 1, the expansion apparatus 100 may include a model generation unit 110 and a calculation unit 120.

The model generation unit 110 determines input data including a past ionospheric delay value 2 associated with a plurality of past time information 1 and a plurality of past time information 1 The target data for the expansion point of the application area can be determined (S12) to generate the model. Here, the expansion time may refer to a time point at which extrapolation to the expansion point of the unavailable area is performed as the extrapolation time point. For example, when real-time extrapolation is performed, the expansion time may refer to the current time (i.e., the current extrapolation time). The model generating unit 110 may determine the past sun / geomagnetic activity information 3 associated with a plurality of past time information 1 as input data when determining the input data (S11).

The model generated by the model generation unit 110 may be a model including input data for SVM (Support Vector Machine) and target data. In addition, the learning through the calculating unit 120, which will be described later, can be performed based on SVM (Support Vector Machine). That is, the input data determined by the model generating unit 110 may be expressed as 'input of SVM' as input data for performing SVM-based map learning.

Specifically, the model generation unit 110 may determine an input for performing SVM-based map learning through an input data determination process (S11). The expansion device 100 can expand the usable area of the ionospheric map by using the functional method. For this purpose, a component having a high correlation with the ionospheric delay can be determined as the input data.

More specifically, the ionospheric delay has a large delay value during the day and a low delay value at night. In addition, the ionosphere delay varies with the season, but spring and autumn have large ionospheric delay values and small latency values in summer and winter. In addition, ionospheric delay is highly correlated with solar activity and geomagnetic activity, and ionospheric delay is large when solar activity and geomagnetic activity are active. In addition, there is a spatial correlation in the ionosphere, but the ionospheric delay value varies greatly depending on the latitude and the ionospheric delay value does not vary widely in the near region. Accordingly, when the above characteristics are all reflected, the ionospheric delay value of the user location where no observing station is present can be estimated by extrapolating. Thus, the expansion device 100 can acquire time information, solar / geomagnetic activity information, Variable elements of the ionosphere delay value can be used.

In this case, the past sun / geomagnetic activity information (3), which is input data, can be used as the solar / geomagnetic activity information and an index indicating the activity of each of the sun and the geomagnetism. For example, the sun / geomagnetic activity information may be F10.7, an index representing sun activity, and the number of sunspots, and Kp (or Ap), an index representing geomagnetic activity. That is, the model generation unit 110 may determine the past values of F10.7, the number of black spots, and Kp as input data (that is, input of the SVM) as past solar / geomagnetic activity information.

Also, the past ionospheric delay value 2, which is input data, may include the past ionospheric delay value for the available area and the past ionospheric delay value for the unavailable area.

Here, the past ionospheric delay value for the usable region may be the ionospheric delay value for the usable region measured at the normal observing station. Or past ionospheric delay values for the available area can be used for the ionospheric delay value of the global ionospheric map provided by the IGS. The expansion device 100 can increase the extrapolation performance to the unavailable area by considering the ionospheric delay value of the available area as the input data (i.e., the input of the SVM) for learning of the ionospheric delay value of the unavailable area. In other words, if the ionospheric delay value of the usable region is included in the extrapolation, extrapolation performance of the ionospheric layer can be enhanced, so that the expansion device 100 can preferably include the ionospheric delay value of the usable region as input data.

Meanwhile, according to the model generating unit 110, an extension point within the unavailable area where the extrapolated expanded ionospheric delay value is to be calculated in the process of determining target data (S12) can be set as a target. The model generating unit 110 may determine the past ionospheric delay value for the unavailable region as input data with respect to the target. Alternatively, the model generating unit 110 may use the ionospheric delay value of the unavailable region as the target data for learning (that is, the SVM target). In this case, considering that the variation of the ionospheric delay value is highly correlated with the time, space, solar activity, and geomagnetic activity, the expansion device 100 may calculate the input of each input (for example, time information, Value, and sun / geomagnetic activity information), it is possible to consider past ionospheric delay values for unavailable regions as input data to learn ionospheric delay values for unavailable regions. The present expansion device 100 can estimate the ionospheric delay of the unavailable region for the expansion point by extrapolating the ionospheric delay based on the learned information from the time when the ionospheric delay value of the unavailable region does not exist, (I. E., Extrapolate the extended ionosphere delay value). ≪ / RTI >

The past ionospheric delay value for the unavailable region may be a post-treatment ionospheric delay value calculated through post-processing based on past observation values at a past time point corresponding to any one of a plurality of past time information (1) And may include the extended ionospheric delay value calculated through the calculation unit 120 with the past time point as the expansion time point.

For example, if there is a post-processed ionospheric delay value, then the post-processed ionospheric delay value may be used as the past ionospheric delay value for the unavailable region. On the other hand, when there is no post-processing ionospheric delay value for a recent predetermined period since an update to the post-treatment ionospheric delay value has been recently made (in other words, a predetermined period has elapsed since the update was made) The extended ionospheric delay value calculated through the calculation unit can be used as the past ionospheric delay value for the region.

Also, the post-treatment ionospheric delay value may be a value calculated using at least one of correction, interpolation and estimation through post-processing based on past observation values. Postprocessed ionospheric delay values may include ionospheric delay values obtained from non-real-time global ionospheric maps provided by post-processing in IGS. In other words, past ionospheric delay values for unavailable areas can be used to interpolate the global ionospheric map provided by the IGS. Also, the post-treatment ionospheric delay value may be calculated from an ionospheric delay value obtained through post-processing of the ionospheric delay value of the available region (for example, a lattice point-based ionospheric delay value derived by inserting / extrapolating the ionospheric delay value of the available region) And the like, but the present invention is not limited thereto.

The calculating unit 120 may perform learning using the model generated by the model generating unit 110 to calculate an extrapolated extended ionospheric delay value with respect to an extension point of the unavailable region at the time of expansion. That is, the calculating unit 120 may perform the learning using the input data generated through the model generating unit 110 and the target data based model (S13). Also, the calculating unit 120 may extrapolate the extended ionospheric delay value to the extension point by extrapolating the extension point based on learning (S14). Here, the time point at which the extrapolation is performed (S14) (i.e., the extrapolation time point at which the extrapolated expanded ionospheric delay value is calculated) may be considered to correspond to the current time point or the expansion time point. The extrapolation performed in the calculation unit 120 may be a real-time extrapolation.

In addition, the plurality of past time information (1) determined as input data in the model generation unit 110 may be information corresponding to a preset past period based on the extended time point. For example, the preset past period may be set to a past one year period based on the extension time point, etc. In this case, the plurality of past time information (1) means time information for the past one year period before the extrapolation time point can do. Here, for example, the setting unit of the period is exemplified as a 'year' unit such as a year, two years, and the like, but the present invention is not limited thereto, and may be set in units of months, weeks, and days. In addition to the plurality of past time information (1), past past ionosphere delay value (2) and past sun / geomagnetic activity information may be information corresponding to a preset past period.

In other words, the input data (SVM input) determined through the model generation unit 110 is input from the past to the extrapolation point (expansion point) before three inputs (i.e., time information, ionospheric delay value, Lt; RTI ID = 0.0 > time-series < / RTI > In an experimental example performed to evaluate the performance of the expansion device 100 to be described later, one year of data before the extrapolation point is used as input data.

On the other hand, when the learning is performed for the first time by the calculating unit 120, the sun / geomagnetic activity index, which is the input data (input of the SVM) used for one learning, is calculated as the ionospheric delay value of the non- The viewpoints can be set to be the same. Generally, extrapolation uses the solar / geomagnetic activity index at the extrapolation point (expansion point), and extrapolation can not be performed if the solar / geomagnetic activity index is not measured at the extrapolation point. However, according to the process of updating (S15) described later, the expansion apparatus 100 updates the past ionospheric delay value of the sun / geomagnetic activity information and the unavailable region to a value immediately before the extrapolation time point, And extrapolation is performed based on the data before the extrapolation time point so that real-time extrapolation is possible.

That is, when the solar / geomagnetic activity information at the expansion time can not be obtained, the calculating unit 120 calculates the solar / geomagnetic activity information of the most recent solar / geomagnetic activity information 2 related to the plurality of past time information Extrapolation can be performed using information as solar / geomagnetic activity information at the time of expansion. Therefore, even if the extension device 100 does not know the solar / geomagnetic activity information at the time when the extrapolation is performed (that is, the expansion point, the current point) when performing the extrapolation, (I. E., Information of the most recent point in the past sun / geomagnetic activity information) to perform extrapolation in real time.

The concrete process of learning through the calculation unit 120 is as follows. After the input data and the target data are determined in the model generation unit 110, the calculation unit 120 may perform the learning process (S13) and the extrapolation process (S14).

The calculating unit 120 may estimate the coefficient of the function based on the SVM using the input data (input of the SVM) and the target data (the target of the SVM) determined by the model generating unit 110. The expansion device 100 can improve the measurement (prediction) performance of the ionospheric delay value by using the SVM having good generalization performance instead of the artificial neural network. In addition, the expansion device 100 can calculate a unique and optimal hyperplane by solving the second optimization problem in the constraint using a Lagrange multiplier based on the SVM.

In solving the nonlinear problem, it is common to use a high-level mapping of the SVM input, and the expansion device 100 can use the kernel trick to map the input of the SVM to a higher dimension. The expansion device 100 may use any one of a Gaussian function, a higher order polynomial function, and a tangent sigmoid function as a kernel function. However, the present invention is not limited thereto, and various known kernel functions can be used. For example, the expansion device 100 can use a Gaussian function as a kernel function. Also, the kernel function applied to the expansion device 100 may have a size of N X N.

As an input used in the learning process S13, for example, an input as shown in Equation 1 may be used.

[Equation 1]

는 시간과 관련된 입력으로서 입력 데이터 중 과거 시간정보(1)에 대한 입력 값일 수 있으며, 사인 함수 또는 코사인 함수로 표현되는 시간 정보와 일 등의 정보가 포함될 수 있다.

는 환경적인 요소와 관련된 입력으로서 입력 데이터 중 과거 태양/지자기 활동 정보(3)에 대한 입력 값일 수 있으며, F10.7, Kp(또는 Ap), 태양흑점 개수 등의 정보가 포함될 수 있다.

는 전리층과 관련된 입력으로서 입력 데이터 중 과거 전리층 지연값(1)에 대한 입력 값일 수 있으며, 가용영역에 대한 과거 전리층 지연값으로서 가용영역의 총전자수(TEC, Total Electron Content) 정보가 포함될 수 있다.here,

May be an input to the past time information (1) of the input data as an input related to time, and may include information such as time information represented by a sine function or a cosine function.

May be input to the past sun / geomagnetic activity information (3) as input related to the environmental factor, and may include information such as F10.7, Kp (or Ap), sunspot number, and the like.

May be an input to the ionospheric delay value 1 of the input data as input related to the ionosphere and may include total electron content (TEC) information of the available area as the past ionospheric delay value for the usable region .

The calculating unit 120 may calculate the kernel function of the input data (SVM input) and then define the secondary optimization problem as shown in Equation (2).

&Quot; (2) "

는 입력 데이터 중

번째 시점과

번째 시점의 입력 데이터로 계산된 커널 함수를 나타낸다.

는 라그랑지 승수를 나타내고,

는 무감도(insensitivity)를 나타낸다.

는 비가용영역에 대한 과거 전리층 지연값으로서 비가용영역의 총전자수(TEC, Total Electron Content) 정보가 포함될 수 있다. 달리 표현하여,

는 외삽 지역(비가용영역의 확장 지점)의 과거 TEC 정보를 포함할 수 있다.here,

Of the input data

And

And the kernel function calculated by the input data at the i-th point.

Represents the Lagrangian multiplier,

Indicates insensitivity.

May include the total electron number (TEC) information of the unavailable area as a past ionospheric delay value for the unavailable area. In other words,

May include past TEC information of an extrapolated region (an extension point of the unavailable region).

값을 계산할 수 있다. 여기서, QP 알고리즘을 구현하는 구체적인 방법으로는 이미 기 알려진 방법들이 적용될 수 있으며, 자세한 설명은 생략하기로 한다.The calculating unit 120 calculates the optimum (k) by using a quadratic programming

The value can be calculated. As a concrete method of implementing the QP algorithm, known methods already known can be applied, and a detailed description will be omitted.

값인

값을 계산할 수 있다. 산출부(120)는 최적의 함수 계수를 계산(또는 추정)하는 학습 과정이 완료된 이후 외삽 수행 과정(S14)을 통해 비가용영역의 전리층 지연값에 대한 외삽을 수행할 수 있다. 이러한 외삽 수행 과정(S14)을 통해 SVM을 기반으로 한 국지적 전리층지도의 가용영역 확장이 이루어질 수 있다. 이때의 외삽식은 하기 수학식 3과 같을 수 있다.The calculating unit 120 calculates the optimal

Value

The value can be calculated. The calculating unit 120 may perform extrapolation on the ionospheric delay value of the unavailable region through the extrapolation process S14 after the learning process for calculating (or estimating) the optimal function coefficient is completed. Through the extrapolation process (S14), the available area of the local ionospheric map based on the SVM can be expanded. The extrapolation equation at this time may be expressed by the following equation (3).

&Quot; (3) "

는 비가용영역의 확장 지점에 대한 외삽된 전리층 지연값(즉, 외삽된 확장 전리층 지연값)을 의미한다.here,

Refers to the extrapolated ionospheric delay value (i. E. Extrapolated expanded ionospheric delay value) to the extension point of the unavailable region.

The expansion device 100 performs a process of updating (or reconfiguring) S15, a learning process S13, and an extrapolation process S13 after the first learning process S13 and the first extrapolation process S14 are performed (S14) can be repeatedly performed to reflect recent ionospheric delay variation.

In other words, the expansion apparatus 100 can perform the learning process (S13) and the extrapolation process (S14) after the model is generated through the input data determination and the target data determination. At this time, the expansion device 100 can perform the real-time extrapolation by repeatedly performing the learning process S13 and the extrapolation process S14 after the initial learning process S13 and the first extrapolation process S14 are performed. At this time, however, the expansion apparatus 100 does not need to generate the model again after performing the first learning process (S13) and the extrapolation process (S14) in performing the real-time extrapolation, (Or reconfiguring) some of them.

After the first execution by the model generation unit 110 and the calculation unit 120, a predetermined time elapses from the first execution time, and a second execution is performed after the elapsed predetermined time as the expansion time The input data in the model generation unit 110 may be updated (or reconfigured) with a plurality of past time information corresponding to a preset past period based on the extension time after the elapsed predetermined time.

For example, when real-time extrapolation is performed, a model generation process (that is, a process of determining input data and target data) through a first process performed by the model generation unit 110 and the calculation unit 120, a learning process S13 ) And the extrapolation process (S14) are performed once for the first time. At this time, the expansion apparatus 100 does not need to newly determine the input data and the target data again when performing the update process S15 after the first execution, The input data and the target data at the oldest point in the data and the target data can be removed (or deleted), and the latest input data and the target data can be added (or updated). This is because, when the size of the kernel function of the SVM used in the expansion device 100 is N X N , if the old data is not deleted, the number of data increases and memory usage and calculation time increase, which is inefficient. Therefore, in order to increase the efficiency of the memory usage and the calculation time in performing the real-time extrapolation, the expansion device 100 performs a process of updating the input data at the old time point and the target data Can be removed.

Here, recent input data and target data may not be data generated (or measured, acquired) at the extension time (current time) at which the extrapolation is performed, but may be data corresponding to the time point before the expansion time (previous time). For example, the recent input data and target data may be data corresponding to the time point t-1, which is the previous time point of the expansion time, instead of the data corresponding to the time point t. Accordingly, the extender 100 can perform extrapolation by performing extrapolation using the sun / geomagnetic activity information before the t point even if the solar / geomagnetic activity information at the time t is not known at the time t when the extrapolation is performed at the current time t, The extended ionosphere delay value can be calculated, and real-time extrapolation is possible.

For example, the expansion device 100 deletes the sun / geomagnetic activity information of the oldest solar / geomagnetic activity information included in the input data through the update process S15 after the first extrapolation, Time of the sun / geomagnetic activity at the most recent time point prior to the time point. In addition to the solar / geomagnetic activity information, the same logic can be applied to the time information and the past ionospheric delay values (including the ionospheric delay value for the available area and the ionospheric delay value for the unavailable area). .

The present expansion device 100 can perform learning based on the SVM using the input data and the target data, and then can extrapolate the ionospheric delay value of the unavailable region. Also, the expansion device 100 can update the input data (input of the SVM) to be used for the real-time extrapolation through the initial learning and extrapolation process (S15). The expansion device 100 can expand the available area of the local ionospheric map in real time through the sequential repetition of the extrapolation process S14 and the update process S15.

Hereinafter, an experiment performed to evaluate the performance of the expansion device 100 will be described with reference to FIGS. 2 to 4. FIG.

According to one experimental example of the present application, we use one year's data from October 1, 2013 to September 30, 2014 to evaluate the extension effect of the local ionospheric map by the expansion device 100 The input data and the target data are determined (or input and target of the SVM are determined) through the model generation unit 110. [ The period is for a large period of solar / geomagnetic activity and to evaluate the extrapolation performance of the local ionospheric map for large periods of ionospheric delay variation.

FIG. 2 shows an example of setting a usable area and a non-usable area for performance evaluation of the available area expansion apparatus 100 of the local ionospheric map according to an embodiment of the present invention.

Referring to FIG. 2, in the experimental example of the present application, the available area is set to a radius of 2,650 km around Daejeon, and the unavailable area is set to a radius of 4,500 km around Daejeon. Extrapolated regions (ie, extension points) are extrapolated to extrapolation and extrapolation distances by 5 °, 10 °, and 15 ° away from the closest available region for each of the four directions Respectively.

In the example of the present invention, SVM-based learning was performed on the input data and the target data determined above. In this case, the kernel function used in learning is Gaussian kernel function, and the total size of kernel function is N X N , and N is 4380.

In the experiment example of the present invention, the first learning is performed, and the extrapolation process and the update (or reconstruction) process are repeated, so that the extrapolated region (that is, the extension point) set one month before October 2014, Delay values were extrapolated and compared to the true values of the IGS global ionospheric map for performance evaluation. The extrapolation error was calculated as the difference between the true value of the IGS global ionospheric map and the extrapolated value (ie, the extrapolated extended ionospheric delay value) by this expansion device (100).

FIG. 3 is a graph illustrating an ionospheric delay value extracted from an IGS global ionospheric map by an extrapolated expanded ionospheric delay value (a graph corresponding to 'SVM extrapolation') by the available area expansion device 100 of a local ionospheric map according to an embodiment of the present invention; (A graph corresponding to 'IGS ionospheric map'). Specifically, FIG. 3 shows the ionospheric delay values and IGS extracted from the IGS global ionospheric map for two days from October 6 to 7, for a predetermined 15 ° North region on the diagram of FIG. 2 100), which is extrapolated to the SVM.

FIG. 4 is a diagram showing the difference between the two ionospheric delay values in FIG. 3 (i.e., the ionospheric delay value calculated by the present expansion device and the ionospheric delay value extracted from the IGS global ionospheric map).

3 and 4, the maximum extrapolation error between the ionospheric delay value calculated by the expansion device 100 and the ionospheric delay value extracted from the IGS global ionospheric map is 0.97 m and the average of the errors is 0.03 m , And the standard deviation was 0.40 m. However, considering that the extrapolation region of the existing extrapolation technique is 5 ° from the nearest observatory, the present expansion device 100 is located at a distance of 15 ° , The spatial distance of the extrapolation is increased (ie, the extrapolation range is increased) as compared with the conventional extrapolation technique. The expansion device 100 can obtain highly accurate extrapolation results for a wider range than the existing extrapolation technique.

Table 1 below shows the results of calculating the root mean square (RMS) value of the extrapolation error for a predetermined extrapolated region on the drawing of FIG. 2 for one month of October 2014.

[Table 1]

Referring to Table 1, it can be seen that all errors up to a 10 ° extrapolation area are less than 0.30 m. For the extrapolated area of 15 °, all except for the southern area, the error was observed to be less than 0.40 m. As a result, extrapolation results with high accuracy over a wider range than the existing extrapolation technique can be obtained through the expansion device 100 as described above.

Hereinafter, the operation flow of the present invention will be briefly described based on the details described above.

5 is a schematic flowchart illustrating a method of expanding a usable area of a local ionospheric map according to an embodiment of the present invention.

The method of expanding the available area of the local ionospheric map shown in FIG. 5 may be performed by the expansion device 100 described above. Therefore, even if omitted in the following description, the contents described with respect to the expansion device 100 can be similarly applied to a description of a method of expanding the available area of the local ionospheric map.

Referring to FIG. 5, in step S51, input data including a plurality of past time information and a past ionospheric delay value associated with the plurality of past time information is determined, and target data for an extension point of a non- Can be determined and a model can be generated.

At this time, in step S51, the input data may include past sun / geomagnetic activity information associated with a plurality of past time information.

Also, in step S51, the past ionospheric delay value may include a past ionospheric delay value for the available area and a past ionospheric delay value for the unavailable area.

The past ionospheric delay value for the unavailable region may be a postprocess ionosphere delay value or a past ionospheric delay value calculated through post-processing based on a past observation value at a past time point corresponding to any one of the plurality of past time information, And may include an extended ionospheric delay value calculated through a calculator with a time point as an extension time point.

For example, if there is a post-processed ionospheric delay value, then the post-processed ionospheric delay value may be used as the past ionospheric delay value for the unavailable region. On the other hand, when there is no post-processing ionospheric delay value for a recent predetermined period since an update to the post-treatment ionospheric delay value has been recently made (in other words, a predetermined period has elapsed since the update was made) The extended ionospheric delay value calculated through the calculation unit can be used as the past ionospheric delay value for the region.

Also, post-treatment ionosphere delay values may include ionospheric delay values obtained from non-real-time global ionospheric maps provided by post-processing in IGS. Also, the post-treatment ionospheric delay value may be calculated from an ionospheric delay value obtained through post-processing of the ionospheric delay value of the available region (for example, a lattice point-based ionospheric delay value derived by inserting / extrapolating the ionospheric delay value of the available region) And the like, but the present invention is not limited thereto.

Also, in step S51, the model may be a model including input data for SVM (Support Vector Machine) and target data.

Next, in step S52, learning using the model generated in step S51 may be performed to calculate an extrapolated expanded ionospheric delay value with respect to the extension point at the extension time.

At this time, in step S52, the learning may be performed based on SVM (Support Vector Machine).

If the solar / geomagnetic activity information at the expansion time can not be obtained, the solar / geomagnetic activity information of the most recent one of the past solar / geomagnetic activity information associated with the plurality of the past time information is stored in step S52. Extrapolation can be performed using the solar / geomagnetic activity information at the expansion point.

In addition, the extension time is the current time, and the extrapolation in step S52 may be a real-time extrapolation.

On the other hand, in step S51, a plurality of past time information may correspond to a preset past period based on the extension time point. If the second execution is performed after the predetermined time elapses from the time point of the first execution after the first execution of the steps S51 and S52 and after the elapse of the elapsed predetermined time, The input data may be updated to a plurality of past time information corresponding to a preset past period based on the extension time after the elapsed predetermined time.

In the above description, steps S51 to S52 may be further divided into additional steps or combined into fewer steps, according to embodiments of the present disclosure. Also, some of the steps may be omitted as necessary, and the order between the steps may be changed.

The method for expanding the available area of the local ionospheric map according to one embodiment of the present invention may be implemented in the form of a program command that can be executed through various computer means and recorded in a computer readable medium. The computer-readable medium may include program instructions, data files, data structures, and the like, alone or in combination. The program instructions recorded on the medium may be those specially designed and constructed for the present invention or may be available to those skilled in the art of computer software. Examples of computer-readable media include magnetic media such as hard disks, floppy disks and magnetic tape; optical media such as CD-ROMs and DVDs; magnetic media such as floppy disks; Magneto-optical media, and hardware devices specifically configured to store and execute program instructions such as ROM, RAM, flash memory, and the like. Examples of program instructions include machine language code such as those produced by a compiler, as well as high-level language code that can be executed by a computer using an interpreter or the like. The hardware devices described above may be configured to operate as one or more software modules to perform the operations of the present invention, and vice versa.

It will be understood by those of ordinary skill in the art that the foregoing description of the embodiments is for illustrative purposes and that those skilled in the art can easily modify the invention without departing from the spirit or essential characteristics thereof. It is therefore to be understood that the above-described embodiments are illustrative in all aspects and not restrictive. For example, each component described as a single entity may be distributed and implemented, and components described as being distributed may also be implemented in a combined form.

The scope of the present invention is defined by the appended claims rather than the detailed description, and all changes or modifications derived from the meaning and scope of the claims and their equivalents should be interpreted as being included in the scope of the present invention.

100: Expandable area of local ionosphere map
110:
120:

Claims (17)

In the method of expanding the available area of the local ionospheric map,
(a) determining input data including a plurality of past time information and a past ionospheric delay value associated with the plurality of past time information, and determining a target data for an extension point of the unavailable region at an expansion time point to generate a model ; And
(b) performing learning using the model, and calculating an extrapolated extended ionospheric delay value with respect to the extension point at the extension time point,
In the step (a), the input data may further include past sun / geomagnetic activity information associated with the plurality of past time information,
The method of claim 1, wherein if the solar / geomagnetic activity information at the extension time can not be obtained, the step (b) further comprises: Wherein the extrapolation is performed using the solar / geomagnetic activity information at the extension time point. delete The method according to claim 1,
The expansion point is the current point,
Wherein the extrapolation in the step (b) is extrapolated in real time. The method according to claim 1,
Wherein, in the step (a), the past ionospheric delay value includes a past ionospheric delay value for a usable area and a past ionospheric delay value for a non-usable area. 5. The method of claim 4,
Wherein the past ionospheric delay value for the unavailable region is calculated by multiplying the posterior processing ionospheric delay value or the past time point calculated through post-processing based on the past observation value at the past time point corresponding to any one of the plurality of past time information And an extended ionospheric delay value calculated through the step (b) as an extension time point. 6. The method of claim 5,
Wherein the post-treatment ionospheric delay value comprises an ionospheric delay value obtained from a global ionospheric map provided in non-real time through post-processing in the IGS. The method according to claim 1,
Wherein, in the step (a), the plurality of past time information corresponds to a preset past period based on the expansion time point,
If a second execution is performed after a predetermined time has elapsed since the first execution of the steps (a) and (b) and after the elapsed predetermined time as an extension time, Wherein the input data of step (a) is updated with a plurality of past time information corresponding to a predetermined past time period based on the extension time after the elapsed predetermined time. The method according to claim 1,
In the step (a), the model is a model including input data and target data for SVM (Support Vector Machine)
Wherein, in the step (b), the learning is performed based on a SVM (Support Vector Machine). In an area expansion device of a local ionospheric map,
Generating a model that determines input data including a plurality of past time information and a past ionospheric delay value associated with the plurality of past time information, determines a target data for an extension point of the unavailable region at an expansion time, part; And
And a calculation unit that performs learning using the model and calculates an extrapolated expanded ionospheric delay value with respect to the expansion point at the expansion time,
Wherein the input data in the model generator further comprises past sun / geomagnetic activity information associated with the plurality of past time information,
Wherein when the solar / geomagnetic activity information at the expansion time can not be obtained, the calculating unit calculates solar / geomagnetic activity information of the most recent one of the past solar / geomagnetic activity information associated with the plurality of past time information, To be used for solar / geomagnetic activity information in the local ionosphere map. delete 10. The method of claim 9,
The expansion point is the current point,
Wherein the extrapolations performed in the calculator are real-time extrapolations. 10. The method of claim 9,
Wherein the past ionospheric delay value in the model generator includes a past ionospheric delay value for the usable region and a past ionospheric delay value for the unavailable region. 13. The method of claim 12,
Wherein the past ionospheric delay value for the unavailable region is calculated by multiplying the posterior processing ionospheric delay value or the past time point calculated through post-processing based on the past observation value at the past time point corresponding to any one of the plurality of past time information And an extended ionospheric delay value calculated through the calculation unit as an extension time point. 14. The method of claim 13,
Wherein the post-treatment ionospheric delay value comprises an ionospheric delay value obtained from a global ionospheric map provided in non-real time through post-processing in IGS. 10. The method of claim 9,
Wherein the plurality of past time information in the model generation unit corresponds to a preset past period based on the extension time point,
When a second execution is performed after a predetermined time elapses from the time of the first execution after the first execution by the model generation unit and the calculation unit and after the predetermined time elapsed after the elapse of the predetermined time, Wherein the input data in the generation unit is updated with a plurality of past time information corresponding to a predetermined past time period based on the extension time point after the elapsed predetermined time. 10. The method of claim 9,
In the model generation unit, the model is a model including input data and target data for SVM (Support Vector Machine)
Wherein the learning is performed on a Support Vector Machine (SVM) basis in the calculation unit. 9. A computer-readable recording medium having recorded thereon a program for executing the method according to any one of claims 1 to 8 on a computer.

Images