JPH11160409A - Position measuring terminal equipment - Google Patents

Abstract

PROBLEM TO BE SOLVED: To reduce the time and cost required for computation, by providing a neural network, etc., which inputs measured values of the position of a position reference station and a propagation time, computs and outputs a measured result of an unknown position. SOLUTION: A receiver 4 of a position measuring terminal 1 receives an electric wave transmitted by a position reference station, and reads and outputs a time of transmission and a position signal written in it. Measured values of the position of the position reference station outputted by a data generation 18b and a propagation time outputted by a physical process computer 22 are inputted to a neural network 16, and it computes and outputs a measured result of an unknown position. The adjustment quantity computer 19c of a learning unit computes and outputs a quantity of adjustment for adjusting the neural network 16 so that the output of the neural network 16 in a learning process may approach a corresponding real value which the data generator 18b outputs. Since it is not necessary to solve a nonlinear equation by taking much time in successive approximation computation, in this way, the time or cost is reduced.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a positioning terminal device that measures a distance from a position reference station using radio waves and obtains an unknown position from the distance.

[0002]

2. Description of the Related Art FIG. 6 shows a global positioning system which is a typical conventional positioning system.
FIG. 2 is a diagram showing the operation of em (hereinafter, referred to as GPS). The operation of the conventional positioning terminal device will be described with reference to FIG.

In FIG. 6, reference numeral 1 denotes a conventional positioning terminal device located at an unknown position and measures the position, 2 denotes a satellite as a position reference station for transmitting radio waves to the positioning terminal device 1, and Is a satellite orbit around which the position reference station 2 orbits the earth.

[0004] FIG. 7 is a Japanese edition of the Geodetic Association of Japan, "Newly revised GPS, precise positioning system using artificial satellites", 198.
9 years, Chapter 5 "Independent positioning" issued by Japan Surveying Association
FIG. 2 is a block diagram showing a configuration of the conventional positioning terminal device 1 shown in FIG.

In FIG. 7, 1 is the conventional positioning terminal device, 4 is a receiver which receives the radio wave transmitted by the position reference station 2 and reads out and outputs the transmission time t _i described therein. Is a clock that outputs the reception time t _{o of the} radio wave, and 6 is the input of the transmission time t _i output by the receiver 4, and the position (x _i , y _i , z _i ) of the position reference station 2.
And outputs the trajectory calculator, 7a, the position location of the reference station _{2 (x i, y i,} z i) and, above the transmission time t _i and the reception time t _o (t _o the terminal clock error [delta] _t ), And the positioning results are converted to rectangular coordinates (x _o , y _o , z _o ),
positioning calculation unit that outputs at [delta] _t, 8a, the orthogonal type the positioning result of the coordinate, the coordinate converter outputs which was converted into latitude and longitude altitude coordinates and 9, the positioning of the longitude and latitude altitude coordinates This is a display to input and display the result. that's all,
The receiver 4, the clock 5, the orbit calculator 6, the positioning calculator 7a, the coordinate converter 8a, and the display 9 constitute the positioning terminal device 1.

Further, in the positioning calculation unit 7a of FIG. 7, a pseudo-distance calculation 10 inputs a measured value τ _i of the propagation time obtained as a difference between the transmission time and the reception time, and outputs a pseudo distance. , A provisional solution (x ′, x ′,
y ′, z ′), δ _t ′, a successive approximation calculator for improving and outputting a tentative solution for the positioning result, and 1
Reference numeral 2 denotes a convergence determiner that inputs the temporary solution output by the successive approximation calculator 11, determines convergence of this value, and outputs the value as a positioning result when the convergence is determined. As described above, the pseudo distance calculator 10, the successive approximation calculator 11,
Then, the convergence determiner 12 configures the positioning calculation unit 7a.

The operation of the above-described conventional positioning terminal will be described below.

First, the position reference station 2 transmits a radio wave indicating the transmission time t _i .

Next, in the positioning terminal device 1, the receiver 4
Receives the radio wave transmitted by the position reference station 2, reads out and outputs the transmission time t _i described therein.

[0010] After the same time as the radio wave is received, or a predetermined time delay, the clock 5 outputs the reception time t _o of the radio wave. However, the clock 5, generally, there is a terminal time error [delta] _t. Thus, wherein the reception time t _o, that contains this terminal time error [delta] _t.

On the other hand, the trajectory calculator 6 receives the transmission time t _i output from the receiver 4. The position reference station 2
Satellite orbit 3 is disclosed, so the transmission time t _i
Of the position reference station 2 at (x _i , y _i , z _i )
Can be obtained by a trajectory calculation considering the transmission time t _i . The trajectory calculator 6 performs this calculation,
The position of the position reference station 2 is output.

The positioning calculator 7a calculates the position of the position reference station 2 output from the orbit calculator 6, the transmission time t _i output by the receiver 4, and the reception time t output by the clock 5. A measurement value τ _{i of the} propagation time obtained as a difference from _o is input. Among them, the measurement value τ _i of the propagation time has a meaning as an apparent propagation time because the reception time t _o includes the terminal clock error δ _t as described above.

Further, in the positioning calculation section 7a, the pseudo distance calculator 10 receives the measured value τ _i of the propagation time and multiplies it by the speed of light, so that the positioning terminal device 1 and the position reference station 2 The pseudo distance r is output.
Means that the pseudo distance r also represents that the terminal clock error [delta] _t affected distance.

The successive approximation calculator 11 is provided with the trajectory calculator 6
Output from the position reference station 2 (x _i , y _i , z
_i ) and the pseudo distance r output by the pseudo distance calculator 10 are input. The relationship between the position of one position reference station 2 (this is referred to as the i-th position reference station 2) and the unknown position where the positioning terminal device 1 is located is represented by the following expression using the measured value τ _i of the propagation time. Is represented. However, it should be noted here that the following equation is expressed by rectangular coordinates. Actually, as the orthogonal coordinates, earth fixed earth-centered orthogonal coordinates called WGS-84 are used.

[0015]

(Equation 1)

In general, the positions of four sets of position reference stations 2 and the pseudo distance to them are input. As a result, the positioning terminal device 1 obtains a certain unknown position in three dimensions and simultaneously obtains the terminal clock error by integrating the four preceding equations.

However, the above equation is a non-linear equation having a square root of the square of an unknown number, and cannot be easily solved. Therefore, the unknown is calculated as a tentative solution (x ′, y ′, z ′),
Expressed as the sum of δ _t ′ and the correction value, the equation is developed for the correction value, and linearization is performed by omitting second-order and higher-order terms under the assumption that the correction value is minute. In this way, the correction value obtained by solving the linear simultaneous linear equation of the obtained correction values is added to the tentative solution, and a new tentative solution, that is, an improved solution is obtained. Then, the unknown is represented again by the sum of the new temporary solution and the new correction value, and the above process is repeated. By this repetition, the tentative solution is gradually improved and approaches a true value. Such a calculation method is called a successive approximation calculation method, and the successive approximation calculator 11 repeatedly performs the successive approximation calculation to improve and output a tentative solution.

The convergence judging unit 12 is provided with the successive approximation calculator 1
1 is input and the convergence is determined. Specifically, the provisional solution is input every time the calculation is repeated, and when the amount of change in the value becomes sufficiently small,
It is determined that they have converged. Then, when it is determined that convergence has occurred, the provisional solution is output as a true solution, that is, a positioning result (x _o , y _o , z _o ), δ _t . on the other hand,
If it is not determined that the convergence has occurred, the provisional solution is again input to the successive approximation calculator 11, and the improvement is repeated. However, as mentioned above. The positioning result output from the convergence determiner 12 is obtained based on the above expression expressed in rectangular coordinates.

The coordinate converter 8a receives the positioning result obtained by the orthogonal coordinates output from the convergence judging unit 12, and converts the positioning result into a local geodetic coordinate system of Japan so as to correspond to a map of Japan. Convert to longitude / latitude coordinates and output. In Japan, the Tokyo Geodetic System (Tokyo
o Datum).

Finally, the display 9 is provided with the coordinate converter 8a.
Inputs the positioning result represented by the longitude / latitude / altitude coordinates output by the system, and displays this position. Generally, it is often displayed on a map.

[0021]

As described above, as described above,
In the positioning, the propagation time of the radio wave from the position reference station 2 to the unknown position is measured, and an unknown position of the positioning terminal device is obtained based on the measured time and the position of the position reference station 2. However, looking at this relationship as a physical process, as shown in FIG. 8, the position of the position reference station 2 and the unknown position are determined first,
The time required for the radio wave to propagate from the position reference station 2 to the unknown position is determined, and a terminal clock error is added to the measured value of the propagation time. This physical process 1
3 is an equation expressing the relationship between the position of the position reference station 2 and the unknown position by the propagation time. However, as shown in FIG.
In, the unknown position is obtained by solving the inverse problem of the physical process from the measured value of the propagation time. The problem here is that, as described above, since the original equation is non-linear, the inverse problem must be solved by iterative calculation with the trouble of the successive approximation method. Further, it is necessary to perform coordinate conversion between a rectangular coordinate system in which this equation is established and a longitude / latitude altitude coordinate system used for association with an actual map. This coordinate conversion is also described in Chapter 7, "How to represent the position on the earth and the coordinate system" in the above-mentioned "Newly revised GPS, precise positioning system using artificial satellites".
As described in detail above, this coordinate transformation is accompanied by the problems of the geocentric coordinate system and the local geodetic coordinate system, resulting in a complicated nonlinear calculation.

From the standpoint of the conventional positioning terminal device, from this standpoint, in order to perform positioning and collate on a map, the inverse problem of the physical process is solved by a successive approximation method, and a coordinate transformation is further calculated. Will be. That is, the conventional positioning terminal device 1 has a problem that the time or cost required for these calculation processes is large.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-mentioned problem, and provides a positioning terminal device capable of performing positioning and collating on a map by processing that does not require additional time or cost. It is intended for.

[0024]

According to a first aspect of the present invention, there is provided a positioning terminal device which inputs the position reference station, the position, and the measured value of the propagation time, and calculates and outputs a positioning result of an unknown position. And a learning unit having an adjustment calculator for outputting an adjustment amount for learning the input / output relationship to the neural network.

According to a second aspect of the present invention, in the positioning terminal device, a data generator that outputs a position of a position reference station and a measured value of a propagation time, and the measured value of the propagation time are input to the learning unit.
A pseudo-range calculator that outputs a pseudo-range, a successive approximation calculator that inputs the position of the position reference station, the pseudo-range, and a tentative solution for the positioning result, and improves and outputs the tentative solution for the positioning result. , A convergence determiner that inputs the temporary solution output by the successive approximation calculator, determines convergence of this value, and outputs the value as a measurement result when convergence is determined. .

According to a third aspect of the present invention, in the positioning terminal device, the learning unit inputs a positioning result, which is an output of the convergence determiner, and converts the result from rectangular coordinates to longitude / latitude altitude coordinates and outputs the result. It is provided with.

According to a fourth aspect of the present invention, in the positioning terminal device, in the learning section, a positioning result, which is an output of the convergence determiner, is directly input to the adjustment amount calculator.

[0028] In a positioning terminal device according to a fifth aspect of the present invention, the learning unit includes: a data generator that outputs a position of a position reference station, an unknown position, and a terminal clock error; And a physical process calculator that inputs a clock error, calculates a measured value of the propagation time, and outputs the calculated value.

[0029] A positioning terminal device according to a sixth aspect of the present invention includes an operation area indicator for outputting a space in which the measurement terminal device is used as an operation area, and the data generator inputs the operation area. In this case, the unknown position is changed and output.

[0030] A positioning terminal according to a seventh aspect of the present invention uses a position reference station whose position is constant or almost constant as the position reference station.

A positioning terminal according to an eighth aspect of the present invention is a position reference station identifier for inputting an output of a receiver and outputting an identification signal of the position reference station, and inputting the identification signal and outputting a position of the position reference station. And a position reference station position readout device.

A positioning terminal device according to a ninth aspect of the present invention includes a position reference station position decoder for receiving an output of a receiver and outputting the position of the position reference station.

[0033]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Some embodiments of a positioning terminal according to the present invention will be described below with reference to the drawings.

Embodiment 1 FIG. 1 is a block diagram showing a configuration of the first embodiment of the present invention. In Figure 1, 15, the position of the position reference station _{2 (x i, y i,} z i)
And a measurement value τ _{i of the} propagation time, and outputs a positioning result in longitude / latitude / altitude coordinates.

In the positioning / coordinate conversion calculating section 15,
Reference numeral 16 denotes a neural network that constitutes the entire positioning / coordinate conversion calculating unit 15 when the positioning is used online.

Further, in FIG. 1, reference numeral 17 denotes a learning unit which uses the neural network 16 as one of the constituent elements when offline.

In the learning unit 17, a data generator 18 a generates a position (x _i , y _i , z _i ) of the position reference station and a measured value of the propagation time τ _i. And an adjustment amount calculator that outputs an adjustment amount for learning the input / output relationship.

Here, the meaning of using the neural network will be briefly described. First, neural networks have historically been a major motivation in research to model animal neural circuits, but here we consider them as just one engineering algorithm. Neural networks are roughly classified into a hierarchical network and an interconnected network. Here, the former refers to the hierarchical network. Furthermore, various types of neural networks of a hierarchical network have been proposed depending on the number of levels in the hierarchy, how to connect units to each other, and the input / output functions of each unit. Regardless of, a neural network of a hierarchical network is generally used.

The feature of the neural network of the hierarchical network when captured as an engineering algorithm is as follows.
This means that it is a mechanism that can learn the input / output functions to be executed when online. For example, in controlling a robot arm, the torque to be controlled at each joint is such that the resulting arm operation is desired. In such a case, when the neural network is used as a control algorithm, the torque is controlled by the neural network when the robot is off-line, and the resulting arm movement and desired arm movement (this is referred to as a “teacher signal”). The neural network is trained by the adjustment signal calculated from. If the learned neural network is used online, the neutral network can control the joint torque that realizes the desired operation of the robot arm. This is, physically,
Inverse function such that the desired robot arm operation is given to the functional relationship resulting from the control of the joint torque and the result of the movement of the robot arm, and the control of the joint torque that realizes it is output. This means that the neural network has been realized.

This provides a means for solving the above-mentioned problem to be solved in the conventional positioning terminal system and the inverse problem of the physical process of obtaining a position from the propagation time of a radio wave. The purpose of the present invention is to solve the above-mentioned problem by realizing an inverse function of a physical process using a neural network.

Next, the operation will be described. The operation includes an online operation related to positioning using the neural network 16 and the neural network 16.
There is an operation at the time of offline for learning. First, an online operation related to positioning will be described.

The operations of the receiver 4, the clock 5, and the trajectory calculator 6 are the same as those of the conventional positioning terminal.

The positioning / coordinate conversion calculation unit 15 calculates the position (x _i , y) of the position reference station 2 output from the orbit calculator 6.
_i , z _i ) and the transmission time t output by the receiver 4
The measured time τ _{i of the} propagation time, which is obtained as the difference between _i and the reception time t _o output by the clock 5, is input.

In the positioning / coordinate conversion calculating section 15,
Neural network 16, the location position of the reference station _{2 (x i, y i,} z i) and the inputs the measured value tau _i of the propagation time, the positioning result of unknown positions corresponding to these values longitude and latitude Calculate and output the value by the altitude coordinate.

Here, the neural network 16
The learning has been completed in advance by the offline operation described below.

Finally, the positioning result is output from the positioning / coordinate conversion calculating unit 15 and is input to the display 9.
The operation of the display 9 is the same as that of the conventional positioning terminal device.

Next, an off-line operation for learning the neural network 16 will be described.

The learning unit 17 uses the neural network 16 as one of its components when offline.

In the learning section 17, the data generator 1
8a, the position of the position reference station _{2 (x i, y i,} z i)
And a measurement value τ _{i of the} propagation time are generated and output.

The positioning calculation section 7b is provided with the data generator 18
a of the position reference station 2 (x _i , y _i , z
_i) and inputs the measured value of the propagation time tau _i. The operation of the positioning calculation unit 7b is the same as that of the positioning calculation unit 7a in the conventional positioning terminal device, and calculates and outputs a correct positioning result in rectangular coordinates by successive approximation calculation.

The coordinate converter 8a inputs a correct positioning result obtained by the rectangular coordinates output by the positioning calculation section 7b. The operation of the coordinate converter 8a is the same as that of the conventional positioning terminal device 1.
And outputs a correct positioning result converted into longitude / latitude / altitude coordinates.

On the other hand, the position (x _i , y _i , z _i ) of the position reference station 2 output from the data generator 18 a and the measured value τ _{i of the} propagation time are simultaneously the constituent elements of the learning unit 17. Is input to the neural network 16. The neural network 16
Calculates and outputs the positioning result in longitude, latitude and altitude coordinates by the same calculation as in the online mode. However, this positioning result in the offline learning process is generally not a correct value.

That is, the position (x _i , y _i , z _i ) of the position reference station 2 output from the data generator 18a and the measured value τ _i of the propagation time are simultaneously obtained by the positioning calculator 7b.
Is input to the neural network 16, and a correct positioning result and a positioning result in a learning process of the neural network 16 are output, respectively.

Then, the adjustment amount calculator 19 inputs the correct positioning result and the positioning result in the learning process of the neural network 16. And the adjustment amount calculator 1
Numeral 9 calculates and outputs an adjustment amount for adjusting the neural network 16 so that the positioning result in the learning process approaches the correct positioning result.

The algorithm for calculating the adjustment signal is an error back propagation method or ErrorBack Propa
This is called a “gating algorithm”. Neural networks are now a widely used methodology, and among them, the algorithm of this back-propagation method is a typical one. No description is given here. The following are easy-to-understand explanations. "Neural Network Information Processing" by Hideki Aso, 1990, published by Sangyo Tosho Co., Ltd.

The adjustment amount output from the adjustment amount calculator 19 is input to the neural network 16. The neural network 16 can gradually output a positioning result closer to a correct positioning result by changing a coupling coefficient between units according to the adjustment amount. A reference value is prepared in advance, and when the difference between the positioning result in the learning process and the correct positioning result becomes smaller than the reference value, the offline learning ends. As a result, the neural network 16 can execute online positioning.

Embodiment 2 FIG. 2 is a block diagram showing a configuration of the second embodiment of the present invention. Embodiment 2
Then, in the learning unit 17, the output of the positioning calculation unit 7b is
The input is directly to the adjustment amount calculator 19b without going through the coordinate converter 8a as in the first embodiment.

Further, the positioning / coordinate conversion calculator 15 which is a component of the first embodiment is replaced by a positioning calculator 7b in the second embodiment.

Next, the operation will be described. Here, the offline operation for learning the neural network 16 will be described first.

In the learning section 17, the data generator 18a
The operation of the positioning calculation unit 7b is the same as that of the positioning terminal device according to the first embodiment.

The adjustment amount calculator 19b inputs the correct positioning result obtained by the rectangular coordinates output by the positioning calculation unit 7b and the positioning result of the learning process output by the neural network 16, and the difference between the two becomes small. Thus, the adjustment amount for adjusting the neural network 16 is calculated and output.

The adjustment amount is input to the neural network 16. The neural network 16
By changing the coupling coefficient between the units according to the adjustment amount, a correct positioning result obtained by the orthogonal coordinates is gradually output. That is, the neural network 16 is learned to output the positioning result of the rectangular coordinates before the conversion into the longitude / latitude / altitude coordinates.

Next, an online operation related to positioning will be described.

The operations of the receiver 4, the clock 5, and the trajectory calculator 6 are the same as those of the conventional positioning terminal device or the positioning terminal device according to the first embodiment.

The positioning calculation unit 7b, like the positioning calculation unit 7a in the conventional positioning terminal device, calculates the position of the position reference station 2 output from the trajectory calculator 6 and the transmission time output from the receiver 4. The measurement value of the propagation time obtained as a difference from the reception time output by the clock 5 is input.

In the positioning calculation unit 7b, the neural network 16 learned offline in advance, like the positioning / coordinate conversion calculation unit 15 in the positioning terminal device 1 according to the first embodiment, stores the position of the position reference station 2 The measurement value of the propagation time is input, and the positioning result is output. However, in the second embodiment, as described above, the positioning calculation learned by the neural network 16 is:
Since the value of the positioning result is represented not by longitude / latitude altitude coordinates but by rectangular coordinates, the positioning calculation unit 7a
Is also a positioning result represented by rectangular coordinates.

Finally, the display unit 9 displays the positioning calculation unit 7a.
Input and output the positioning result represented by the rectangular coordinates output by. In particular, the positioning terminal device according to the second embodiment is effective when displaying the positioning result in relation to the distance from another arbitrary point. For example, in the second embodiment, the display 9 obtains and displays a range of a radius of 10 km from the position of the positioning terminal device obtained as a positioning result.

Embodiment 3 FIG. 3 is a diagram illustrating the operation of the third embodiment of the present invention and the fourth embodiment described below.

In FIG. 3, reference numeral 20 denotes a fixed position,
Alternatively, it is a position reference station which is almost constant. Such a position reference station 20 is realized, for example, by an airship stopped in the air.

FIG. 4 is a block diagram showing a configuration of the third embodiment of the present invention.

In the learning section 17 shown in FIG.
Is an operation area indicator that outputs a space in which the positioning terminal device 1 is used as an operation area, and 18b is an input of the operation area output by the operation area indicator 21, and the position and longitude of the position reference station 20 The data generator 8b for outputting the unknown position (x _o , y _o , z _o ) represented by the latitude and altitude coordinates and the terminal clock error δ _t , 8b comprises an unknown position represented by the longitude and latitude altitude coordinates. A coordinate converter for inputting a position, converting the position into rectangular coordinates, and outputting the position; and a position (x _i , y _i , z _i ) of the position reference station 20 and an unknown position converted into the rectangular coordinates This is a physical process calculator that inputs (x _o , y _o , z _o ) and terminal clock error δ _t , calculates and outputs a measured value τ _i of the propagation time.

Further, in FIG. 4, reference numeral 23 denotes a position reference station identifier for inputting the output of the receiver 4 and outputting the identification result of the position reference station 20, and 24 stores the position of each position reference station 20. Each position reference station position storage unit 25, which outputs the position in response to the inquiry, receives the identification result of the position reference station 20 output by the position reference station position reader 23, and stores the position reference station 20 corresponding to the identification result. Is a position reference station position readout unit which inquires and inputs the position of each position reference station position storage unit 24 and outputs the same.

Next, the operation will be described. First, an online operation related to positioning will be described.

First, the position reference station 20 whose position is constant or almost constant transmits a radio wave indicating the transmission time t _i . Further, an identification signal for identifying the position reference station 20 is added to this radio wave. This signal may be a message or a spread spectrum code.

Next, in the positioning terminal device 1, the receiver 4
Receives the radio wave transmitted by the position reference station 2, reads out the transmission time t _i and the identification signal described therein, and outputs the read out signal.

[0076] After the same time as the radio wave is received, or a predetermined time delay, the clock 5 outputs the reception time t _o of the radio wave. However, the clock 5, generally, there is a terminal time error [delta] _t. Thus, wherein the reception time t _o, that contains this terminal time error [delta] _t.

On the other hand, the position reference station identifier 23 receives the identification signal output from the receiver 4 and outputs a result of identification as to which position reference station 20 the received radio wave is from.

Each position reference station position storage 24 stores the position of each position reference station 20 and outputs the position in response to an inquiry.

Then, the position reference station position reader 25
Inputs the identification result of the position reference station 20 output from the position reference station identifier 23, and stores the position of the position reference station 20 corresponding to the identification result in each of the position reference station position storage units 2.
4 and input it to output the position of the position reference station 20.

The positioning / coordinate conversion calculating section 15 calculates the position of the position reference station 20 outputted by the preceding position reference station position reader 25, the transmission time t _i outputted by the receiver 4 and the time outputted by the clock 5. The measurement value τ _{i of the} propagation time obtained as a difference from the reception time t _o is input.

In the positioning / coordinate conversion calculating section 15, similarly to the positioning / coordinate conversion calculating section 15 in the positioning terminal device 1 according to the first embodiment, the neural network 16 learned in advance offline is connected to the position reference station 20. Enter the position and the measured value of the propagation time,
Outputs the positioning result represented by longitude, latitude and altitude coordinates.

Finally, the display 9 inputs the positioning result represented by the longitude / latitude / altitude coordinates output from the positioning / coordinate conversion calculating section 15 and displays it.

Next, only the offline operation for learning the neural network 16 will be described.

The learning unit 17 uses the neural network 16 as a component when it is offline. The learning unit 17 calculates the relationship between the value of the position of the position reference station 20 and the value of the positioning result corresponding to the measured value of the propagation time, and causes the neural network 16 to learn the relationship as a function. This point is the same as that of the positioning terminal device according to the first or second embodiment, but the method for obtaining the relationship is based on the calculation in accordance with the physical process shown in FIG. 8, and is shown in FIG. It differs from the positioning terminal device according to the first or second embodiment in that the calculation is not for solving the inverse problem.

In the learning section 17, first, the operation area indicator 21 specifies a space in which the positioning terminal device 1 is used, and outputs the space as an operation area. This means that the neural network 16 is trained only in the space where the positioning terminal device 1 is actually used in the range of the domain and the range of the function, thereby avoiding learning of data having no operational significance. There is. For example, there is no case where a positioning terminal device is carried at a position at an altitude of minus 1,000 meters, that is, at the ground floor to perform positioning. Also, this operation area specification
It would be more natural to consider that the coordinates should be given in longitude and latitude altitude coordinates suitable for human interpretation and operation of social systems, rather than rectangular coordinates such as geocentric coordinates.

Next, the data generator 18b receives the operation area output from the operation area indicator 21, inputs an unknown position (true value) in the operation area, the position of the position reference station 20, and the terminal. Outputs the clock error. That is, the domain of the unknown position (true value) does not need to cover the entire combination of numerical values, but only has to cover the range of the operation area.

The coordinate converter 8b calculates the position (true value) of the position reference station 20 given by the longitude / latitude altitude coordinates.
Is input from the data generator 18b, and a result obtained by converting the data into rectangular coordinates is output.

The physical process calculator 22 calculates the position (x _i , y) of the position reference station 20 output from the data generator 18b.
_i , z _i ), the terminal clock error δ _t , and the unknown position (true value) represented by the rectangular coordinates output by the coordinate converter 8b. Then, a radio wave is transmitted from the given position reference station 20, received at a given unknown position (true value), and this propagation time is determined by a reception time measured by a clock having a given terminal clock error. Calculate the measured value along this actual physical process,
Output as a measured value of the propagation time.

The neural network 16 inputs the position of the position reference station 20 output from the data generator 18b and the measured value of the propagation time output from the physical process calculator 22. The neural network 16 outputs not only the positioning result but also the estimated value of the terminal clock error obtained at the same time when the learning is offline.

The adjustment amount calculator 19c inputs the unknown position (true value) output from the data generator 18b and the terminal clock error, and simultaneously obtains the positioning result output from the neural network 16 in the learning process. And the estimated value of the terminal clock error. Then, an adjustment amount for adjusting the neural network 16 so that the output of the neural network 16 in the learning process approaches the corresponding true value output by the data generator 18b is calculated and output. I do.

The adjustment amount is input to the neural network 16. The neural network 16
Changes the coupling coefficient between the units according to the adjustment amount, so that a correct positioning result represented by longitude / latitude / altitude coordinates and a correct terminal clock error are gradually output.

Embodiment 4 FIG. FIG. 5 is a block diagram showing a configuration of the fourth embodiment of the present invention.

In FIG. 4, reference numeral 26 denotes the position reference station 2
This is a position reference station position decoder which receives the output of the receiver 4 for receiving radio waves from 0 and outputs the position of the position reference station 20.

Next, the operation will be described. Since the offline operation for learning is the same as that of the third embodiment,
Only the online operation related to positioning will be described.

First, the position reference station 20 whose position is constant or almost constant transmits a radio wave indicating the transmission time. Further, a position signal indicating the position of the position reference station 20 is added to this radio wave.

Next, in the positioning terminal device 1, the receiver 4
Receives the radio wave transmitted by the position reference station 2 and reads out and outputs the transmission time and the position signal described therein.

Next, the position reference station position decoder 26 receives the position signal output from the receiver 4, decodes the position of the position reference station 20 that has transmitted the received radio wave from the position signal, and outputs it. . This output is input to the positioning / coordinate conversion calculation unit 15.

The positioning / coordinate conversion calculator 15 and the display 9
Is similar to that of the third embodiment.

[0099]

According to the first aspect of the present invention, it is not necessary to solve a nonlinear equation by taking the time and effort of successive approximation calculation, so that it is possible to perform positioning with reduced time or cost load.

According to the second aspect of the present invention, since the value of the positioning table is determined by the same calculation processing as that of the conventional positioning terminal device, the positioning result is the same as that of the conventional positioning terminal device, and Thus, the continuity of the output is guaranteed even when the switching operation is performed between the two.

According to the third aspect of the present invention, not only the trouble of the successive approximation calculation for solving the non-linear equation but also the problems of the geocentric coordinate system and the local geodetic coordinate system are added, and the quadrature which becomes a complicated non-linear calculation is added. The time or cost burden in the process of converting the coordinates from the coordinates to the longitude / latitude coordinate system can also be reduced.

According to the fourth aspect, the positioning result can be output in a format suitable for calculating the distance from an unknown position where the positioning terminal is located to an arbitrary position. It is suitable for calculations such as obtaining a range of a radius of 10 km from the position of.

According to the fifth aspect, since the positioning table is created by calculation in accordance with the physical process, there is no approximation error caused by the successive approximation method, and an accurate value can be obtained. It is possible to save the trouble of repeated calculations. Furthermore, since it is not necessary to take a form for solving the inverse problem, it is not necessary to handle a physically impossible input combination, and it is possible to handle only a necessary input combination.

According to the sixth aspect of the present invention, in learning the neural network, only the combination of inputs that are practically meaningful for the user of the positioning terminal device can be learned, so that the time required for learning is shortened. Further, it is possible to prevent the learning accuracy of meaningful data from deteriorating due to the learning of meaningless data.

According to the seventh aspect, the load for performing the trajectory calculation can be eliminated, and deterioration of the positioning accuracy due to the calculation error can be prevented.

According to the eighth invention, each position reference station can use its position for positioning only by adding its identification signal to the radio wave.

According to the ninth aspect, even when the position of the position reference station is not truly constant, it can be used for positioning by successively informing the position.

[Brief description of the drawings]

FIG. 1 is a first embodiment of a positioning terminal device according to the present invention;
FIG. 3 is a block diagram showing the configuration of FIG.

FIG. 2 is a second embodiment of the positioning terminal device according to the present invention;
FIG. 3 is a block diagram showing the configuration of FIG.

FIG. 3 is a third embodiment of the positioning terminal device according to the present invention;
And FIG.

FIG. 4 is a third embodiment of the positioning terminal device according to the present invention;
FIG. 3 is a block diagram showing the configuration of FIG.

FIG. 5 is a fourth embodiment of the positioning terminal device according to the present invention;
FIG. 3 is a block diagram showing the configuration of FIG.

FIG. 6 is a diagram showing the operation of GPS, which is a typical conventional positioning system.

FIG. 7 is a block diagram showing a configuration of a conventional positioning terminal device.

FIG. 8 is a diagram showing a physical process in which a measured value of a propagation time is determined.

FIG. 9 is a diagram showing, as an inverse problem, a process of positioning calculation when positioning is performed from a measured value of a propagation time.

[Explanation of symbols]

Reference Signs List 1 positioning terminal device, 2 position reference station (satellite), 3 satellite orbits, 4 receiver, 5 clock, 6 orbit calculator, 7 positioning calculator, 8 coordinate converter, 9 display, 10 pseudo distance calculator, 11 successive Approximation calculator, 12 convergence determiner, 13
Physical process, 14 Positioning calculation process (inverse problem), 15
Positioning / coordinate transformation calculation unit, 16 neural network, 17 learning unit, 18 data generator, 19 adjustment amount calculator, 20 position reference station (airship), 21 operation area indicator, 22 physical process calculator, 23 position reference station identifier , 24 each position reference station position storage device, 25 position reference station position reading device, 26 position reference station position decoding device.

Claims (9)

[Claims] 1. A positioning terminal device for determining an unknown position by measuring a propagation time of a radio wave with a position reference station, inputting the position of the position reference station and the measured value of the propagation time, and A positioning terminal comprising: a neural network that calculates and outputs a positioning result of a position; and a learning unit having an adjustment calculator that outputs an adjustment amount for learning the input / output relationship of the neural network. apparatus. 2. A data generator that outputs a position of a position reference station and a measured value of a propagation time to the learning unit, a pseudo distance calculator that inputs the measured value of the propagation time and outputs a pseudo distance, A successive approximation calculator for inputting the position of the position reference station, the pseudo distance, and a provisional solution for the positioning result, improving and outputting the provisional solution for the positioning result, and the provisional solution output from the successive approximation calculator. , A convergence determiner that determines the convergence of this value, and outputs the value as a measurement result when the convergence is determined,
The positioning terminal device according to claim 1. 3. The learning unit according to claim 1, further comprising a coordinate converter for inputting a positioning result output from the convergence determiner, converting the result from rectangular coordinates to longitude / latitude / altitude coordinates, and outputting the result. The positioning terminal device according to claim 2. 4. The positioning terminal device according to claim 2, wherein in the learning unit, a positioning result, which is an output of the convergence determiner, is directly input to the adjustment amount calculator. 5. A data generator that outputs a position of a position reference station, an unknown position, and a terminal clock error to the learning unit, and the position reference station, the unknown position, and the terminal clock error are input and propagated. The positioning terminal device according to claim 1, further comprising: a physical process calculator that calculates and outputs a measured value of time. 6. An operation area indicator for outputting a space in which a measurement terminal device is used as an operation area, wherein the data generator inputs the operation area and changes the unknown position within the operation area. The positioning terminal device according to claim 5, wherein the positioning terminal device outputs the data. 7. The positioning terminal device according to claim 1, wherein a position reference station whose position is constant or substantially constant is used as said position reference station. 8. A position reference station identifier for receiving an output of a receiver and outputting an identification signal of the position reference station, and a position reference station position reader for receiving the identification signal and outputting the position of the position reference station. The positioning terminal device according to claim 7, wherein the positioning terminal device is provided. 9. The positioning terminal device according to claim 7, further comprising a position reference station position decoder for receiving an output of a receiver and outputting a position of said position reference station.