JP2880449B2 - Moving object monitoring method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method of monitoring a moving object applied to, for example, air traffic control, and is used when the motion model of the moving object is unknown on the observation side or even when the motion model is known. In the case of lack of certainty or when the input of information from the sensor that measures the position of the moving object is delayed, a motion model of the moving object is estimated in real time using the information from the sensor, and the position of the moving object is estimated. The present invention relates to a moving body monitoring method for displaying a course, a speed, and the like.

[0002]

2. Description of the Related Art FIG. 9 is a block diagram showing a configuration of a tracking apparatus using a conventional moving object monitoring method disclosed in, for example, Japanese Patent Application Laid-Open No. 7-218611.
1 is a moving object group including a plurality of moving objects, 2 is a sensor group including a plurality of sensors, 11 is a sensor number indicating which sensor is output from data sent from the sensor group 2,
A sensor data processing device for assigning the number of the moving body and the reception time, a filter processing device for estimating the position and speed of the moving body from data output from the sensor assuming a plurality of motion models of the moving body in advance, Reference numeral 13 denotes a target estimating processor for selecting a motion model having the smallest error among the motion models calculated by the filter processor, 14 an input / output processor, 15 an input device, and 9 an output device.

Next, the operation will be described. When the motion model of a moving object captured by a conventional sensor is unknown, a method called an αβ tracker is generally used for tracking the moving object. In this method, weighted averaging is performed on the predicted position and the input position using the smoothing parameter α on the position, assuming that data on the position of the moving object, which is the tracking target, is input at almost regular intervals. , A weighted averaging process is performed on the predicted speed and the input speed using the smoothing parameter β to estimate the smoothed position and the smoothed speed. The values of the smoothing parameters α and β are dynamically changed according to the reliability of the tracking target. Further, as this modification, means for setting the smoothing parameters α and β separately in the traveling direction of the moving body and the direction perpendicular to the traveling direction, and means for introducing the γ parameter by incorporating the acceleration element have been proposed and put into practical use. Has been This technique is
In general, since the calculation is performed using a simple formula, the calculation load of the information processing apparatus is not required large, and the method is suitable for simultaneously calculating a large number of moving objects in real time. In addition, although there is a calculation means such as a least square method which requires a long calculation time, it is not generally used as a method for simultaneously tracking a plurality of targets because a calculation load is applied while accurately tracking a moving object.

[0004] A tracking device using the conventional moving object monitoring method disclosed in Japanese Patent Application Laid-Open No. 7-218611 shown in FIG. 9 is disclosed in Japanese Patent Application Laid-Open Nos. 61-184377 and 61-195381. Based on the disclosed “target tracking filter”, the sensor group 2
And a means for inputting data from a plurality of sensors, calculating a trajectory of a moving object as a tracking target based on a plurality of motion models, and selecting a motion model with the least error. As shown in FIG. 9, a moving object is detected by one or a plurality of sensor groups 2, and the sensor data processing device 11 performs standardization so that data input from each sensor can be commonly processed. Further, based on the standardized data, the tracking of the moving object is performed by the motion model assumed by the filter processing device 12 and the calculation formula. The result of the tracking is sent to the target estimation processing device 13, and the target estimation processing device 13 selects one of the motion models having the least error. In addition, the data of each exercise model is displayed on the output device 9 through the input / output processing device 14 and, based on the judgment of the operator who sees the data, the intention of the operator is selected from the input device 15 through the input / output processing device 14 and the target estimation processing device 13 A means for allowing the operator to intervene in the tracking by notifying the operator of the tracking is disclosed.

When there are a plurality of sensors such as radars for tracking a moving object, the conventional moving object monitoring method uses the most reliable tracking information (tracking information) among the tracking information obtained from each sensor in the sensor overlapping area. (High quality) is used as the system data. JP-A-4-208891 and JP-A-5-1
In the multiple radar automatic tracking processing means disclosed in Japanese Patent Laid-Open No. 200001, when tracking information having the highest tracking reliability is adopted as system data, that is, when the radar is switched, the tracking reliability is reduced among the plurality of sensors. In order to reduce the fluctuation of the trajectory caused by the change of the rank, a means for adding a condition for switching is used.

In the above-described conventional method for monitoring a moving object, a sequential calculation method is used in which the position, speed, acceleration, and the like of each moving object are sequentially calculated even when a plurality of moving objects are tracked. .

By the way, a radar is a typical sensor for a moving object, and a reflected wave or a response wave signal from the moving object is received by a plurality of pulses in response to an interrogation wave from a conventional radar, and a response time is obtained. Sensors and GPs with high accuracy using monopulse angle measurement technology, in addition to the primary and secondary radars that calculate the position from the center angle of the pulse and the center of the pulse
Practical use of automatic subordinate monitoring technology that measures the position, speed, and course of the moving object itself using techniques such as S, transmits the information to the moving object monitoring device that monitors the moving object, and displays the information on the moving object monitoring device Is about to be achieved. The automatic dependent monitoring can also be regarded as a sensor in a broad sense in that information such as positional information of the moving body is obtained.

[0008] In the transition period of such technology, it is considered that there is a long period in which a high-precision sensor and a conventional sensor coexist. In addition, in the case of a high-precision sensor, since it passes through a plurality of relay devices, it takes time for the information to reach the mobile monitoring device at the final destination, or the mobile has its own observation device such as a position. In some cases, transmission may not be possible to the mobile monitoring device at the final destination at necessary and sufficient intervals due to radio wave allocation. Therefore, in some cases, data from a variety of sensors whose measurement accuracy, reception interval, and time delay of data reception vary depending on the type of sensor are comprehensively processed, and the optimal solution of the trajectory of the moving object is obtained. There are situations where you have to find out.

[0009]

Since the conventional moving object monitoring method is configured as described above, simply using or deforming the αβ tracker method does not always provide sufficient tracking performance. When the speed is changing and the course is changing, there is a problem that the tracking ability is not sufficient and the tracking is slow. Although such a conventional moving object monitoring method has been unavoidable in order to track a large number of moving objects, a recent increase in the calculation performance / price ratio requires more calculation time, but the accuracy is high. It has become possible to employ good tracking means such as the least squares method.
Therefore, it is conceivable to apply a parallel processing processor having a good calculation performance / price ratio to such a moving object monitoring method. However, in the αβ tracker method, since the smooth prediction calculation is individually performed each time data of the moving object is input from the radar, there is a problem that the calculation efficiency does not increase even if the parallel operation processor is used.

When an analog video of a radar signal itself that has not been digitally processed by the radar and a tracking target that has been digitally processed are superimposed and displayed on a display device,
There is a problem in that the display of the tracking target is delayed by the time for performing the digital processing and the tracking processing, and the display positions of the tracking targets do not match for a fixed time. Furthermore, if data from a plurality of sensors having different data reception times are displayed as they are, there is a problem that a difference occurs between digital data and an error occurs in a distance between moving objects displayed at a certain moment.

[0011] Further, in the multiple radar automatic tracking processing means disclosed in JP-A-4-2088891 and JP-A-5-100021 as means for comprehensively processing information from a plurality of sensors, the tracking is most performed. It merely describes the means for selecting a highly reliable sensor and the means for switching the sensor, and there is no consideration to comprehensively process the data of various sensors as described above. There was a problem that a solution could not be found.

Further, the target tracking filter disclosed in Japanese Patent Application Laid-Open Nos. 61-184377 and 61-195381 requires that the position, velocity and acceleration are known as basic data used for tracking calculation. Since there is no consideration for data when the speed and acceleration are unknown, and there is no consideration for comprehensively processing the data of a variety of multiple sensors, the optimal solution for the trajectory of the moving object is also found. There was a problem that it was not possible.

Further, in a moving object monitoring method disclosed in Japanese Patent Application Laid-Open No. Hei 7-218611, which includes a step of comprehensively processing information from a plurality of sensors, data is input from the plurality of sensors and a plurality of targets are input. The means of calculating with the motion model and selecting the motion model with the least error is shown, but only by calculating each of the existing motion models and selecting the optimal motion model There is no disclosure of how to comprehensively process data from various types of sensors, and similarly, there has been a problem that an optimal solution of the trajectory of the moving object cannot be found. Furthermore, even if the operator's intervention is permitted, by using means for individually providing complicated tracking-related information to the operator and asking for the operator's judgment, even if a small number of tracking targets are possible, a large number of tracking targets can be simultaneously obtained. There is a problem that it is necessary to track and monitor, for example, it is unsuitable for providing information to an operator for air traffic control, so that reliable air traffic control cannot be performed.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-described problems, and comprehensively processes data from a variety of sensors having different characteristics to realize accurate tracking of a moving object. It is another object of the present invention to provide a moving object monitoring method capable of incorporating parallel arithmetic processing suitable for high-speed processing of a parallel arithmetic processor in tracking trajectory calculation.

Further, according to the present invention, by displaying the trajectories of a plurality of moving objects on the display device at the same time, it is possible to reduce the error in the distance between the moving objects due to the time zone shift, and to reduce the radar signal and the like. It is an object of the present invention to obtain a moving object monitoring method capable of accurately displaying a superimposed display with analog data.

A further object of the present invention is to provide a moving object monitoring method capable of estimating a deviation of unknown sensor-specific size in real time and correcting the tracking trajectory calculation.

[0017]

Means for Solving the Problems A mobile monitoring method according to this invention, when the type of sensor that outputs position information in polar coordinates, of the common coordinate system for determining the trajectory about the horizontal position of the movable body The observation formula for the X coordinate is multiplied by a weight inversely proportional to the distance in the Y direction of the local coordinate system of the sensor,
The observation formula for the Y coordinate is X in the local coordinate system of the sensor.
The trajectory of the moving object is estimated by multiplying the distance in the direction by a weight inversely proportional to the distance.

According to the moving body monitoring method of the present invention, when position information is input from a sensor that outputs position information in polar coordinates, the time at which the position information was measured is estimated from the azimuth included in the position information. Is what you do.

The moving object monitoring method according to the present invention evaluates the past speed fluctuation of the moving object based on the expression of the moving object trajectory obtained by the calculation of the least square method, thereby obtaining the expression of the moving object trajectory. The exercise model to be adopted is determined.

The mobile object monitoring method according to the invention, wherein the trajectories of evaluating the variation in speed of the moving body are those wherein motion models represented by the quadratic function formula with respect to time.

The mobile object monitoring method according to the invention, the result of the evaluation of the variation of the past velocity of moving body, the moving body in a past predetermined period with respect to time as the expression of the trajectory if continues to uniform motion Select the motion model represented by the primary box formula,
If you keep changing the speed, select a motion model expressed by a quadratic function with respect to time as the trajectory formula.If you start changing the speed from constant velocity motion, use the quadratic function with respect to the latest time as the trajectory formula. Select the motion model represented by the equation, or, if you move from a state where the speed is continuously changed to constant velocity motion, select the motion model represented by a linear function with respect to the latest time as the trajectory equation. is there.

According to the moving object monitoring method of the present invention, the formula of the trajectory of the moving object is calculated based on the least squares method based on the position information from the first sensor which outputs the position information including an error whose size is unknown. And the second sensor, which is more accurate than the first sensor,
The error is estimated by applying the measured time of the position information input from the sensor or the time estimated to be measured to the estimated trajectory formula.

According to the moving object monitoring method of the present invention, the trajectory of the other moving object is estimated by using the least squares method based on the position information of the other moving object from the first sensor, and the estimation is performed. The equation of the trajectory of the other moving object is corrected in consideration of the error.

According to the moving object monitoring method of the present invention, each time position information is obtained from the second sensor, the moving object obtained by using the least square method based on the position information from the first sensor is obtained. The error is re-estimated from the equation of the trajectory, the error estimated this time and the error estimated in the past are weighted and averaged, and the value is newly estimated as an error.

The mobile object monitoring method according to the present invention, in displaying the position of the plurality of mobile display device, by applying the display time in the equation of the locus estimated respectively to a plurality of mobile The position is obtained.

The moving object monitoring method according to the present invention includes a fixed amount of previously obtained position information for a plurality of moving objects that need to estimate the trajectory formula using the least square method, and each moving object includes: On the other hand, a trajectory calculation table to which areas of the same size are allocated is provided, and dummy data is written into cells of the trajectory calculation table where no position information exists.

The mobile object monitoring method according to the invention, each cell of the trajectories calculation table includes the weight for the sensor accuracy of the sensors to obtain the position information, the weights in the dummy data is written cells That is zero.

The mobile object monitoring method according to the invention, with temporarily storing data including the position information inputted from a sensor of several, in each cell there is no data provided tracking table dummy data is written Each time the trajectory formula for a plurality of moving objects is estimated, (the number of moving objects for which the trajectory formula should be estimated) once stored in the tracking table ×
The oldest data of the same amount as the latest data amount (the maximum number of data accumulated for each moving body since the previous trajectory estimation) is deleted from the trajectory calculation table, and the latest data is tracked. The data is transferred from the table to the locus calculation table.

According to the moving object monitoring method of the present invention, when the speed and acceleration measured by the moving object are input, the speed and the acceleration are obtained from the trajectory formula estimated for the moving object, and these are measured by the moving object itself. The speed and acceleration measured by the moving object itself are corrected by weighting and averaging the obtained speed and acceleration.

[0030]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS One embodiment of the present invention will be described below. Embodiment 1 FIG. FIG. 1 is a block diagram schematically showing a mobile monitoring apparatus for realizing a mobile monitoring method according to Embodiment 1 of the present invention. In the figure, the same parts as those in FIG. Is omitted. In FIG. 1, reference numeral 3 designates a single / secondary radar input from each sensor of the sensor group 2 such as a primary / secondary radar, an SSR mode S radar, or a device for receiving information such as position information and speed observed by a moving object. Alternatively, a target input device that receives input data such as position information of a moving object group 1 including a plurality of moving objects, sensor types, and the like;
A correlation processing device having a function of searching for existing tracking targets corresponding to the input data of the moving object group 1 input from the sensors and establishing a correspondence, and a tracking table 5 having a two-dimensional structure. The moving object identification information for identifying the moving object such as a beacon code or a mode S address of each tracking target, a parameter indicating the reliability of tracking, the type of the trajectory formula finally selected as a result of the trajectory calculation, and A plurality of tracking attribute data 1 to n indicating the attributes of the tracking target, such as the value of the coefficient of the trajectory equation, position information of the tracking target, which is data input from each sensor of the sensor group 2 at one time, and its position The position information acquisition time and the like at which the information was obtained and the sensor accuracy coefficient (weight related to sensor accuracy) sk corresponding to the type of sensor are set to 1
A plurality of input data sets 1 to 3 are sequentially registered by accumulating information of a certain format as a set of input data sets by the number of times input for the tracking target since the previous trajectory calculation.
m are stored. Each of the input data sets 1 to m mainly composed of data input from the sensor group 2 other than the tracking attribute data 1 to n indicating the attribute of the tracking target is used every time the trajectory calculation using the least square method is performed. As will be described later, the data is transferred to another table for calculation, and is therefore cleared from the tracking table 5. Note that the tracking table 5 has columns for a predetermined maximum number of tracking targets.

In FIG. 1, reference numeral 6 denotes a smoothing prediction processing control device for controlling the smoothing prediction calculation of the trajectory using the least square method, and 7 denotes a parallel calculation processing provided with a parallel operation processor for performing the smoothing prediction calculation. A device 71 is a trajectory calculation table provided in the local memory of the parallel calculation processing device 7. The trajectory calculation table 71 is stored in the tracking table 5, transferred from the tracking table 5, and sequentially input from the sensor group 2. , Ie, position information of the tracking target, and information of a fixed format in which one set of the position information acquisition time or the like at which the position information was obtained and the sensor accuracy coefficient sk according to the sensor type are divided into a certain number of sets. It is a two-dimensional table similar to the tracking table 5 provided. However, tracking table 5
Unlike the contents of the trajectory calculation table 71, the contents of the trajectory calculation table 71 are not cleared even when the trajectory calculation is performed, and each time an input data set having a predetermined capacity is transferred from the tracking table 5, the oldest input data sets are deleted. Updated to the latest multiple input datasets.

Further, in FIG. 1, reference numeral 8 denotes a mobile unit which receives a result of a trajectory calculation for tracking by the parallel calculation processing unit 7 and performs display control according to the purpose of the mobile unit monitoring apparatus according to the first embodiment. The information processing device 9 is a display device that displays a trajectory calculation result for tracking. An arithmetic processing unit 10 includes a target input device 3, a correlation processing device 4, a tracking table 5, a smoothing prediction processing control device 6, and a mobile information processing device 8. Note that the arithmetic processing unit 10 may be configured by being divided as appropriate.

Next, the operation will be described. As shown in FIG. 1, in the moving object monitoring device according to the first embodiment provided on the ground, each sensor of the sensor group 2 detects the horizontal position of the moving object in the moving object group 1, and the position information is obtained. Output to the target input device 3. Depending on the type of sensor in the sensor group 2, information such as a beacon code, a mode S address, or a mobile-body-specific code is added. Further, data output from each sensor of the sensor group 2 may include data including altitude information and data not including the altitude information. Is performed. Therefore, in the following operation, the trajectory calculation relating to the horizontal position of the moving body group 1 is performed. In addition, depending on the type of the sensors in the sensor group 2, if it takes time to detect the horizontal position of the moving object and transfer the data to the target input device 3, time stamp (time stamp) information is output to the target input device 3. The horizontal position of the moving object is measured in addition to the data to be obtained, and the obtained horizontal position obtaining time is reported. If it does not take much time to detect the horizontal position of the moving object and transfer the data to the target input device 3, the time when the target input device 3 obtains the data is the time when the position information included in the data is measured, that is, the position information. It can be estimated that it was the time of acquisition. It is assumed that the position information has an error dispersed around the true position.

When a plurality of data relating to each mobile unit in the mobile unit group 1 are input from each sensor of the sensor group 2, it is clear which sensor the data belongs to. The corresponding sensor accuracy coefficient sk is determined by referring to a table or the like that defines these values in advance according to the type of sensor.

As described above, data such as position information of a plurality of moving bodies is input to the target input device 3 from a plurality of sensors of the sensor group 2 asynchronously. When the coordinate system of each sensor of the sensor group 2 is different for each sensor, the target input device 3 converts the coordinate system into a unified coordinate system predetermined by the moving body monitoring device.

The target input device 3 calculates the acquisition time of the horizontal position of the moving object from time stamp information added to data input from the sensors of the sensor group 2 or by calculation using the direction of the moving object with respect to the sensor. You can know. In the latter case, i.e., when receiving data from a sensor that rotates at a constant period and has means for obtaining the position information of the moving object from the azimuth angle at which the position information was obtained, the detection processing time and the target input device If the transfer time up to 3 can be easily estimated to be shorter than a certain time, the target input device 3
Calculates the azimuth of the moving object with respect to the sensor from the position information (position coordinates) of the moving object and estimates the time of obtaining the position information of the moving object.
It is assumed that the rotation cycle of the sensor is known in advance, and that the time is transmitted when the rotating body points to the reference direction.

Hereinafter, the details of the above-described estimation processing of the position information acquisition time of the moving object in the moving object monitoring method will be described with reference to the flowchart shown in FIG. Target input device 3
Obtains data on the moving object in the moving object group 1 from one sensor of the sensor group 2 in step ST201,
In step ST202, it is determined whether or not the sensor is of the type of the rotating body sensor, and it is possible to estimate the position information obtaining time at which the position information was obtained from the azimuth. If the sensor cannot estimate the position information acquisition time, the target input device 3 determines in step ST203
, It is determined whether or not the data on the moving object includes time stamp information. When the data on the moving object includes the time stamp information, the target input device 3 estimates that the time stamp information is the position information obtaining time in step ST204. On the other hand, when the input data on the moving object does not include the time stamp information, the target input device 3 discards the data on the moving object in step ST205 because the position information obtaining time is unknown.

On the other hand, in the determination of step ST202, it is determined that the acquired data on the moving body is of the type of the rotating body sensor and that the position information obtaining time can be estimated from the azimuth of the moving body with respect to the sensor. In this case, in step ST206, the target input device 3 determines whether the obtained data on the moving object has polar coordinates. Generally, in the case of a rotating body sensor, position information (coordinates) of a moving body is obtained from an antenna rotation angle and a response time of a radio wave, so that polar coordinates are source data. In this case, if the data is sent to the moving object monitoring device in the form of polar coordinates, it is difficult to calculate and display the trajectory by using only the polar coordinates, so that the data is converted into rectangular coordinates and processed, but the data of the polar coordinates is separately utilized. Therefore, it is possible to obtain the position information acquisition time of the moving body from this. However, it is conceivable that the computer on the sensor side converts the polar coordinates into rectangular coordinates and then sends the rectangular coordinates to the moving object monitoring apparatus. In such a case, it is necessary to convert the polar coordinates again to obtain the azimuth. Therefore,
If the acquired data on the moving object does not have polar coordinates, the target input device 3 determines in step ST207 that
The position information represented by the rectangular coordinates is converted into a polar coordinate format. Then, the target input device 3 executes step ST20.
In step 8, the acquisition time of the data on the moving object acquired from the azimuth coordinates in the polar coordinates, that is, the position information acquisition time, is calculated by equation (1).

Location information acquisition time = NT + SC × AZ / 360 (1)

Here, NT is the reference azimuth passing time (time when the rotating body points to the reference azimuth), SC is the rotation cycle time of the sensor, and AZ is the azimuth angle (in degrees) of the moving body.

The target input device 3 performs processing in accordance with the flowchart shown in FIG. 2 on the obtained data on the moving body, and then includes data such as position information, position information obtaining time, and sensor accuracy coefficient sk. The input data set and other information (moving object identification information and the like) input from the sensor are transmitted to the correlation processing device 4.

The correlation processing device 4 determines whether the input data set relating to the newly acquired moving object corresponds to any of the existing tracking targets or a new tracking target (correlation processing). The input data set relating to the moving object input first and other information (moving object identification information, etc.) input from the sensor all become the input data set 1 and the tracking attribute data of the tracking target as they are, and the input data input thereafter. These input data sets, including the set, are used for calculating the trajectory of the moving object such as the direction and speed. The correlation processing device 4 stores the input data set relating to the moving object input from the target input device 3
The moving object identification information for identifying the moving object such as a newly input beacon code and the beacon code of the registered tracking target, etc. Is determined by comparing with the moving object identification information. In addition, instead of the moving object identification information such as a beacon code, whether or not the predicted position of the tracking target and the position information included in the newly input data set are similar to each other may be related to the input moving object. It is possible to determine which of the existing tracking targets registered in the tracking table 5 the input data set corresponds to.

In the correlation processing of the moving object by the correlation processing device 4, an input data set having abnormal position information is discarded. On the other hand, if it is determined that the input data set relating to the input moving object is one of the existing tracking targets registered in the tracking table 5, the data is registered in the tracking table 5. On the other hand, when it is not possible to determine which tracking target registered in the tracking table 5 is the data, the correlation processing device 4 inputs the data from the input data set and the sensor regarding the obtained moving body. Other information (moving object identification information and the like) is registered as a new tracking target input data set and tracking attribute data, or the input data set and the like are discarded.

FIG. 3 is an explanatory diagram showing a specific example of the tracking table 5. FIG. 3A shows the entire configuration of the tracking table 5, and FIG. 3B shows the contents of one input data set. As described above, each column of the tracking table 5 includes a movement for identifying a moving object such as a mode A beacon code (by the secondary radar) or a mode S address (by the SSR mode S radar) of each tracking target. Multiple tracking indicating the attributes of the tracking target, such as body identification information, parameters indicating the reliability of tracking, information identifying the equation of the motion model finally selected as a result of the trajectory calculation, and the value of the coefficient of the equation of the motion model Attribute data 1 to n and a plurality of data input at one time from each sensor of the sensor group 2, that is, time stamp information and position information (common coordinate system X coordinate, common coordinate system Y
Coordinate, the absolute value of the X coordinate of the sensor coordinate system, the absolute value of the Y coordinate of the sensor coordinate system, the presence / absence of altitude information, the altitude information) and the sensor accuracy coefficient sk according to the sensor type. A plurality of input data sets 1 to m are stored in which information is accumulated for the number of times input for the tracking target since the previous trajectory calculation and registered sequentially in cell units. Further, as shown in FIG. 3B, in an invalid cell in each column of the tracking table 5, that is, in an invalid input data set, the sensor accuracy coefficient sk is zero, and the other information has no meaning. This is filled with dummy data (usually zero) indicating this.

As is well known, when an interrogation wave is transmitted from a secondary radar, which is a sensor, and the aircraft has a transponder corresponding to the interrogation wave, the aircraft transmits a mode A beacon code for individually identifying itself to the secondary radar. Send to radar. Also, when an interrogator is transmitted from the SSR mode S radar developed from the secondary radar and the aircraft has a transponder corresponding to the interrogator, the aircraft transmits a mode S address for individually identifying itself to the SSR mode S radar. Send to Further, in the future, a case may be considered in which the aircraft itself sends a mobile unit identification code for identifying the aircraft. The moving object identification information for identifying the moving object such as the mode A beacon code or the mode S address is registered as the tracking attribute data in the tracking table 5 when the mobile monitoring device first observes the tracking target. You.

A parameter indicating the reliability of tracking registered in the tracking table 5 as one tracking attribute data for each tracking target is a parameter determined in the process of calculating a trajectory for tracking. For example, if the data of the tracking target is reliably obtained from the sensors of the sensor group 2 every time, and new data having position information of a position close to the position predicted by the trajectory calculation is input, the tracking target is tracked. Can be determined to be highly reliable. If it is determined that the tracking reliability is high, the search range can be reduced when correlating the data registered in the tracking table 5 with the newly input data. There are various other methods for determining the reliability of tracking, but the details are omitted.

The sensor accuracy coefficient sk is determined in advance for each sensor type. Specifically, the sensor accuracy coefficient sk is, for example, proportional to the average error of the position of the sensor, that is, the reciprocal of the standard deviation. When the sensor accuracy coefficient sk of a sensor having poor measurement accuracy is set to 1, the measurement accuracy ( The sensor accuracy coefficient sk of a sensor whose triple is the reciprocal of the standard deviation) is three. The sensor accuracy coefficient sk is not limited to this, but also depends on how measurement accuracy is defined.

Further, in the case of a rotating body type sensor as shown in FIG. 4, the position information of the moving body output from the sensor generally includes an error δθ in the azimuthal direction of the polar coordinates. In this case, if this error is converted to a rectangular coordinate system, it becomes proportional to the distance r. Therefore, it is necessary to multiply the sensor accuracy coefficient sk by an accuracy coefficient (inversely proportional to the distance) taking this error into account.

FIG. 5 is a flowchart showing the trajectory calculation processing in the smoothing prediction processing control device 6 and the parallel calculation processing device 7. The following description will be made with reference to this flowchart.
The trajectory calculation processing is performed at regular intervals in accordance with the update timing of the display data of the display device 9.

First, in step ST301, all the input data sets input from the sensor group 2 that have been stored in the tracking table 5 since the previous trajectory calculation are stored in the trajectory calculation stored in the local memory of the parallel calculation processing device 7. The number of data sets accumulated most among all the tracking targets is multiplied by the maximum number of tracking targets so as to be transferred to the table 71, so that the parallel calculation processing device 7 has a data capacity equal to the table capacity. The old input data set of the locus calculation table 71 in the local memory is shifted. The oldest input data set overflowing from the trajectory calculation table 71 is discarded. The data amount of the oldest input data set discarded is equal to the data amount of the latest input data set to be transferred. Then, the smoothing prediction processing control device 6
In step ST302, the input data set that has been input from the sensor group 2 and accumulated in the tracking table 5 during the period until the next trajectory calculation is performed on the vacant trajectory calculation table 71 is transmitted. As described above, the data amount to be transmitted is equal to the data amount of (the maximum number of tracking target data sets in which the most data sets are accumulated × the maximum tracking target number). Dummy data is packed in a data area of the trajectory calculation table 71 where no input data set exists, that is, a cell.

FIG. 6 is an explanatory diagram of the trajectory calculation table 71 showing the state of the above-described transfer processing. FIG. 6 shows a state in which the latest plural input data sets are transferred from the tracking table 5 shown in FIG. Is shown. The data input from the sensor group 2 is accumulated as shown in FIG. 3 during the period until the next trajectory calculation is performed in the local memory of the parallel calculation processing device 7 on the moving body monitoring device side. Assuming that the number of input data sets that have been accumulated most in are three tracking targets 3, the maximum number of tracking targets is N.
As shown in FIG. 6, an invalid input data set is included (3
Input data sets for (row × N columns) are transferred from the tracking table 5 to the trajectory calculation table 71. As described above, cells in which no input data set exists in the trajectory calculation table 71 are filled with dummy data as in the tracking table 5.

Next, in step ST303, the smoothing prediction processing control device 6 sets the X-axis and Y-coordinates of the horizontal position of the moving object as primary variables using time as variables in order to estimate the trajectory of the moving object. Alternatively, based on a plurality of input data sets stored in the trajectory calculation table 71, assuming that the function can be represented by a quadratic function, the coefficients of the functions are calculated for both the primary function and the quadratic function. Create an equation for The primary box formula assumes the case of a motion model in which the moving body is traveling straight at the same speed without changing the course, and the secondary box formula assumes that the moving body changes the course or changes the speed. The case of an exercise model that exercises is assumed. In the trajectory calculation in the moving object monitoring method according to the first embodiment, it is assumed that the trajectory of the moving object is represented by a motion model of a primary function or a secondary function, and data obtained by observing the coefficients of the respective equations (input data set include)
Is calculated using the least-squares method so that the sum of the squares of the differences between the two and the values calculated from the equations of the trajectories of these motion models is minimized.

In the case of the quadratic function formula, X of the horizontal position of the moving body
The coordinates are represented by equation (2). X (t) = a _x2 × (T−t) ² + v _x2 × (T−t) + w _x2 (2)

Here, T is the target time, and t is the time of obtaining the position information of the moving object obtained from an arbitrary sensor of the sensor group 2. Note that the target time is the current time (the time when the collection of the input data is temporarily delimited for the trajectory calculation), the display time (the time when the predicted position is displayed on the display device 9), or another time. This is a time according to the purpose of the mobile monitoring apparatus. Normally, several steps are taken from the calculation of the trajectory to the output to the display device 9, so that a certain delay occurs until the display depending on the design. Therefore,
By setting the target time as the display time in consideration of the delay time, it is possible for the operator to see the most real-time position on the display device 9. However, in this case,
Since the prediction factor becomes larger, the accuracy is lower than in the case of the current time. Which one to use depends on the design concept and the needs of the user. Another example of the target time is a real-time update time at which the display of the moving object by the analog data is changed in accordance with the rotation of the radar when the analog data is superimposed and displayed at the same time.

If the values a _x2 , v _x2 , w _x2 closest to the motion model expressed by the above equation (2) can be obtained by the calculation of the least squares method, the horizontal position of the moving body in the X direction can be predicted.
The speed and acceleration of the moving body in the X direction can be estimated. When the position information obtaining time t1 and the X coordinate X1 of a certain tracking target obtained from an arbitrary sensor of the sensor group 2 are applied to a motion model that can be expressed by the equation (2), the observation equation

X1 = a _x2 × (T−t1) ² + v _x2 × (T−t1) + w _x2 (3)

Is obtained. Similarly, the X coordinate of another position information acquisition time of the same sensor with respect to the same tracking target, or the X coordinate of another position information acquisition time of another sensor is (t2, X2),..., (Tn, Xn). Then, the following observation formulas of the motion model expressed by the above formula (2) to which the least square method is applied are obtained.

[0058]

(Equation 1)

By the way, when predicting the trajectory of a moving object,
Taking into account speed and course changes, data closer to the current time is more important than older data. That is, if the data at the past time and the data at the latest time are treated equally, if the trajectory is changed recently, the data may be pulled by the past data and the error may increase. On the other hand, there is a case where the amount of data is insufficient due to recent data alone and a correct trajectory cannot be derived. Therefore, in consideration of such a point, the observation data is multiplied by the time-related weight, ie, the time-related accuracy coefficient τk, which is set to have a larger value as the current time is closer to the current time as given by the following equation. Also makes it possible to minimize the overall error of the coefficients of the trajectory equation.

Τk = F (T−t) (5) where t is the time at which the observation formula was obtained, that is, the position information obtaining time, T is the target time, and the function F (x) is the value of x F (x) = 1 / (x + C), where C is a constant, or F (x) = (T-tc-x) , Where tc is a constant. Which formula is used is determined according to the characteristics of the mobile monitoring device.

Next, in step ST304, the smoothing prediction processing control device 6 adds the sensor accuracy coefficient sk, rotation, which depends on the sensor accuracy described above, to the observation formulas (3) and (4) obtained as described above. A normal equation is set by weighting the accuracy coefficient considering the accuracy of the body type sensor and the accuracy coefficient τk regarding time, and the most probable values of the coefficients a _x2 , v _x2 , and w _x2 are obtained by the least square method. As is well known, the least squares method is an equation that minimizes the square of the difference between the observed value and the value calculated by the equation to which the least squares method is applied (in this case, the trajectory equation). The equation is determined by finding the coefficient of. Therefore, if the meaning of each observation value is equal, there is no particular need to weight, but as described above, data that is close to the current time is more important than old data when speed and course changes are taken into account. . Furthermore, when acquiring data from sensors with different accuracy as assumed in the present invention,
If the data with high precision is given a certain weight, a trajectory formula closer to the truth can be obtained. Therefore, by multiplying each observation formula of the motion model to which the least squares method is applied by a sensor accuracy coefficient sk which is proportional to the measurement accuracy of the sensor, the accuracy is doubled as compared with the sensor having the reference measurement accuracy. The data from the sensor having the measurement accuracy as a reference is one observation of the sensor data having the measurement accuracy, but is observed four times (as described above, the value of the least square method is the same as the observed value. This is because the coefficient of the equation is determined such that the square of the difference from the value calculated by the equation of the trajectory to which the least-squares method is applied is minimized), and the trajectory calculation is executed. That is, in the calculation of the least squares method for determining the coefficient of the trajectory equation, the sensor accuracy coefficient s
By weighting k (≧ 1), it is assumed that the data has been observed more. The same applies to the accuracy coefficient in consideration of the accuracy of the rotator-type sensor. As shown in FIG. 4, the error is proportional to the distance from the sensor. In the observation formula relating to the X coordinate of the system, the weight zyk = 1 / (syk + sc), which is inversely proportional to the value obtained by adding the constant sc to the distance syk in the Y direction of the local coordinate system of the sensor.
Is further multiplied by the sensor accuracy coefficient sk, and the observation formula for the Y coordinate includes the distance sxk in the X direction of the local coordinate system of the sensor.
Weight zxk = 1 / in inverse proportion to the value obtained by adding a constant sc to
(Sxk + sc) is further multiplied by the sensor accuracy coefficient sk. If the sensor is not a rotating body type, zx
k and zyk are constant values corresponding to the accuracy of the sensor.

Here, the following equation (6) is defined in order to obtain the most probable values of the coefficients a _x2 , v _x2 , w _x2 by the least square method. In equation (6), the weight such as the accuracy coefficient τk relating to time is squared and included in each equation. However, this may be changed and set to other than square ( 6) This is a problem of how to define the weights of the equations and the like, and may of course be fixed to the square.) How to set it is a matter of adjustment.

[0063]

(Equation 2)

Using the equation (6) defined as described above, the following three-dimensional linear simultaneous equation is established between the coefficients a _x2 , v _x2 and w _x2 .

K _x1 × a _x2 + K _x2 × v _x2 + K _x3 × w _x2 = K _x4 (7) K _x2 × a _x2 + K _x3 × v _x2 + K _x5 × w _x2 = K _x6 K _x3 × a _x2 + K _x5 × v _x2 + K _x7 × w _x2 = K _x8

As described above, if the Y coordinate of the horizontal position of the moving object is represented by a quadratic function using time as a variable,

Y (t) = a _y2 × (T−t) ² + v _y2 × (T−t) + w _y2 (8)

The following three-dimensional simultaneous linear equation is similarly established. _{K y1 × a y2 + K y2} × v y2 + K y3 × w y2 = K y4 (9) K y2 × a y2 + K y3 × v y2 + K y5 × w y2 = K y6 K y3 × a y2 + K y5 × v y2 + K _y7 × w _y2 = K _y8

As described above, the primary box formula assumes the case of a motion model in which the moving body travels straight at a constant speed without changing the course. The trajectory in the case of assuming a primary function in the same manner as in the case of the secondary function can be expressed as follows.

X (t) = v _x1 × (T−t) + w _x1 (10) Y (t) = v _y1 × (T−t) + w _y1 (11)

Regarding the X coordinate of the horizontal position of these moving bodies, a system of linear equations K _x3 × v _x1 + K _x5 × w _x1 = K _x6 (12) K _x5 × v _x1 + K _x7 × w _x1 = K _x8

[0072] is established a binary simultaneous linear equations K with respect to the Y-coordinate _{y3 × v y1 + K y5 ×} w y1 = K y6 (13) K y5 × v y1 + K y7 × w y1 = K y8

Is established. In the same manner as described above, the smoothing prediction processing control device 6 limits the number of observation formulas (ie, the number of data used for trajectory calculation) to data close to the current time (the number of data is a parameter). A three-dimensional linear simultaneous equation and a two-dimensional linear simultaneous equation for obtaining the coefficients of the trajectory equation in the case of a motion model that can be expressed by the following quadratic function formula and linear function formula are established.

X (t) = a _x4 × (T−t) ² + v _x4 × (T−t) + w _x4 (14) Y (t) = a _y4 × (T−t) ² + v _y4 × (T− t) + w _y4 (15) X (t) = v _x3 × (T−t) + w _x3 (16) Y (t) = v _y3 × (T−t) + _{wy 3} (17)

The number of data as a parameter is predetermined according to the purpose of the mobile monitoring apparatus. In general,
If a lot of past data is used in the trajectory calculation, the trajectory from the past to the present can be more accurately estimated. However, the moving object does not always move on the same trajectory, and may change direction midway. In such a case,
Being particular about past data will adversely affect future trajectory predictions. Therefore, the number of data is determined in advance in consideration of the characteristics of the mobile monitoring device and the characteristics of the tracking target (observation cycle, tracking target speed, turning capability, turning frequency, etc.).

The calculation processing in the trajectory calculation of the smoothing prediction processing control device 6 and the parallel calculation processing device 7 is executed collectively for a plurality of tracking targets by using vector calculation. When the vector calculation is performed, the same calculation is performed for all the columns of the trajectory calculation table 71 shown in FIG. 6, that is, for all the tracking targets capable of performing the trajectory calculation in row units. In parallel calculation processing using vector operation, it is preferable to avoid decision logic by branching as much as possible. For this purpose, the following means can be mentioned.

(1) The trajectory calculation is executed assuming that all the tracking targets are effective. Even if the effective total tracking target number or data number for which the trajectory calculation can be performed is reduced, the calculation load is not reduced, but the maximum processing performance defined by the specifications of the mobile monitoring device can be suppressed.

(2) In order to make the assumption of (1) valid, even if invalid data is calculated, the trajectory calculation result is not adversely affected. That is, the sensor accuracy coefficient sk of the invalid data in the trajectory calculation table 71 shown in FIG. 6 is set to zero, and is used to obtain each coefficient of the above-described three-dimensional linear simultaneous equation or the two-dimensional linear simultaneous equation. 6) There is no adverse effect on the calculation as in the equation.

(3) As already described with reference to FIGS. 3 and 6, the number of valid data in each column of the tracking table 5 is not the same, but is different. The data of (effective data number) × (maximum number of tracking targets) is transferred to the trajectory calculation table 71 provided on the local memory of the parallel calculation processing device 7 for vector calculation.

(4) Before transferring data of (the maximum number of accumulated valid data) × (the maximum number of tracking targets) to the trajectory calculation table 71 on the local memory of the parallel calculation processing device 7 every time the trajectory is calculated. Then, a process of shifting the old input data set of the tracking target in the local memory is executed for the vector operation.

(5) In the least-squares method calculation, the largest calculation load is the calculation of the sum of the three-dimensional linear simultaneous equations or the two-dimensional linear simultaneous equations as shown in equation (6). is there. In the calculation of these sums,
The calculation is performed using the same formula regardless of invalid data.
Further, for example, a motion model that can be expressed by a quadratic function equation limited to data close to the current time as shown in Expression (14) is represented by X
Coefficient K ' _x1 of the ternary system of equations when applied to coordinates
To K ′ _x8 and K ′ _{y1 to} K ′ _y8 are coefficients K _{x1 to} K _x8 in equation (7) and coefficients K _{y1 to} K in equation (9), respectively.
Similar to _y8 , it is calculated based on equation (6). Therefore, if the coefficients K _{x1 to} K _x8 and K _{y1 to} K _y8 are sequentially calculated from the latest data, the vector calculation is performed up to the specified number of rows of the trajectory calculation table 71 in FIG. Branch until the coefficients K ' _{x1 to} K' _x8 and K ' _{y1 to} K' _y8 of the ternary linear equations are applied when a motion model that can be expressed by a quadratic function limited to data close to the current time is applied. Trajectory calculations can be performed without the need. Then, after temporarily storing the calculation results of the coefficients K ′ _{x1 to} K ′ _x8 and K ′ _{y1 to} K ′ _y8 , the calculation of the sum is continued and the coefficient K shown in equation (7) is obtained.
The coefficient K _y1 ~K _y8 shown in _x1 ~K _x8 and (9) can be calculated.

(6) Since the solution of each simultaneous equation uses the solution of an ordinary simultaneous equation using a determinant, the calculation formula is omitted here, but each solution is obtained in the form of a fraction of A / B.
Therefore, when calculating each solution, it is necessary to perform a judgment process in order to avoid division by zero.
, The vector operation can be continued without branching until the stage of calculation of the denominator and the numerator immediately before obtaining each solution of the above eight simultaneous linear equations. That is, a series of calculations of the coefficients of the eight simultaneous linear equations and the numerator and denominator of each floor are continuously executed without branching by postponing the division for obtaining each solution.

Thus, the X coordinate and Y of each tracking target
After a series of calculations of the coefficients of the four simultaneous linear equations for each of the coordinates and the numerator and denominator of each floor,
The smoothing prediction control unit 6 finds solutions to eight simultaneous linear equations for each tracking target. If the denominator (B) of each solution is zero or the number of valid data is generally small, an impossible solution will occur. No movement if the number of valid data is 1,
When the number of valid data is 2, only the linear expression of the trajectory is used. In each of the motion models that can be expressed by the primary function formula and the quadratic function formula, the obtained constants (w _x1 , w _x2, etc.) represent the position of each tracking target at the target time T, and the coefficient of the variable primary term (−v _x1 , −v _x2, etc.) represent the speed of each tracking target, and the coefficients (2 × a _x1 , 2 × a
_x2 ) represents the acceleration of each tracking target.

Next, the smoothing prediction processing control device 6 performs (2)
Based on the quadratic function formulas of the X coordinate and the Y coordinate shown in the equations (8) and (8), it is evaluated how the speed estimated by the calculation of the least squares method has fluctuated in the past. It is determined which of the four simultaneous linear equations for each of the X coordinate and the Y coordinate is selected as the equation of the trajectory of each coordinate.

First, in step ST 305, the smoothing prediction processing control device 6 evaluates the equation (2) and (2) for each tracking target in order to evaluate the past fluctuation of the speed estimated by the calculation of the least squares method. 8) Using the values accumulated in the past for the coefficients a _x2 and a _y2 of the quadratic terms of the quadratic function of the quadratic function of the formula for the predetermined number of times, the trajectory of the primary function was calculated for each of these coefficients. In the same manner, a primary function equation of coefficients a _x2 and a _y2 is obtained. These coefficients are _expressed as A _X2 (m), A
_{Assuming Y2} (m), the primary box formula is

[0086] _{A X2 (m) = q x} × m + r x (18) A Y2 (m) = q y × m + r y (19)

Can be represented by the following format. As in the case of the above-described trajectory calculation, the parallel calculation processing device 7 performs a vector operation without branching until the stage of calculation of the denominator and the numerator immediately before obtaining each solution of two simultaneous linear equations relating to these two linear function equations. Can be continued.

Further, in step ST306, the smoothing prediction processing control device 6 selects a tracking target and proceeds to step S306.
At T307, it is determined whether or not the tracking target is valid (whether or not it is a target for which trajectory calculation is to be performed). If it is determined that the tracking target is valid, the horizontal target of the tracking target is determined at step ST308. Using the parameters p ₁ and p ₂ preset by the moving object monitoring device to evaluate the past fluctuation of the estimated speed, the following cases are classified based on which trajectory formula is used in the X coordinate of the position. Judge according to.

(1) | q _x | <p ₁ and | r _x | <
For p _2, in this case, the tracking target is made uniform motion with respect to the X-axis direction, the variation of the velocity estimates not, as an expression of the trajectory of the X-coordinate of the tracked target (10) formula X (t) = v _x1 × (T−t) + w _x1 is adopted.

(2) | q _x | <p ₁ and | r _x |>
For p _2, in this case, estimates that the tracking target is under speed variation with respect to the X-axis direction, X of the above equation (2) as an expression of the trajectory of the X-coordinate of the tracked target (t) = a _x2 × ( Tt)
² + v _x2 × (T−t) + w _x2 is adopted.

(3) When | q _x | ≧ p ₁ and | a _x2 (latest value) | <p ₂ , in this case, the tracking target has changed from a speed change state to a constant speed movement state in the X-axis direction. And X (t) = v _x3 × (T−t) + w _x3 in the above equation (16) is adopted as the equation of the trajectory of the X coordinate of the tracking target.

(4) When | q _x | ≧ p ₁ and | a _x2 (latest value) | ≧ p ₂ , in this case, the tracking target is in a state in which speed fluctuation has started from a constant velocity motion state in the X-axis direction. And X (t) = a _x4 × (T−t) ² + v in the above equation (14) as an equation of the trajectory of the X coordinate of the tracking target.
_x4 × (T−t) + w × ₄ is adopted.

Further, in step ST309, the smoothing prediction processing control device 6 sets the Y position of the horizontal position of the tracking target.
In the coordinates, it is similarly determined which locus formula is to be adopted. In step ST310, the information indicating the trajectory formulas adopted for the X and Y coordinates of the effective horizontal position of each tracking target and the values of the coefficients of the trajectory formulas are stored in the tracking table 5 in the tracking table 5. Register as target tracking attribute data. Therefore, the registered tracking attribute data includes the position information at the target time T.

[0094] The smoothing prediction processing control unit 6 determines in step ST
In step 311, the above-mentioned step ST
It is checked whether the processing from 306 to ST310 has been performed, and if there is any remaining tracking target, the process proceeds to step ST3.
Returning to 06, the processing of steps ST306 to ST310 is similarly repeated. Once all tracking targets have been processed,
In step ST312, the smoothing prediction process control device 6 determines the value of the coefficient of the trajectory equation calculated in the trajectory calculation for all valid tracking targets, that is, the coordinates of the horizontal position,
The speed and the like are transmitted to the mobile information processing device 8. Further, in step ST313, the smoothing prediction processing control device 6 clears the data area in the tracking table 5 in which the input data sets 1 to m input from the sensors of the sensor group 2 are stored. Before clearing this data area, the tracking table 5 is prepared for recovery from the failure of the mobile monitoring apparatus.
A method for preserving the contents of the present invention can be considered, but is omitted here because it is not directly related to the object of the present invention.

The moving body information processing apparatus 8 calculates the value of the coefficient of the trajectory equation calculated in the above trajectory calculation for all effective tracking targets and other data according to the purpose of the moving body monitoring apparatus (such as a radar signal). Display data at the target time by synthesizing the
Send to In addition, the mobile object information processing device 8 predicts a future position using the value of the input coefficient, and monitors the result of the prediction for abnormal approach between the mobile objects, monitoring of entry into the mobile object entry prohibited area, It is used for estimating the arrival time of a point to be set. Further, if necessary, the mobile information processing device 8 creates display data for this purpose and transmits it to the display device 9. The display device 9 displays the transmitted display data and provides the information to the monitor. Therefore, the relative position between the moving bodies can be accurately displayed at the same target time, and the moving body position can be accurately displayed with analog data such as a radar signal.

As described above, according to the first embodiment, the sensor measurement is performed on data obtained from a plurality of sensors having different characteristics such as measurement accuracy and data input time to the mobile monitoring device. Multiply the sensor accuracy factor sk set in consideration of the accuracy by the observation formula of the motion model applying the least squares method, and observe the accuracy factor τk related to the time set so that the data closer to the current time has a larger value. By multiplying the expression, the data can be comprehensively processed to realize accurate tracking, and the effect of monitoring the approach of the moving object and predicting the target arrival time can be obtained early and accurately. Furthermore, since data transfer, arithmetic processing, and the like for effectively achieving parallel arithmetic processing in the parallel arithmetic processing device 7 are adopted, an effect that trajectory calculation can be executed at high speed can be obtained.

In the first embodiment, the sensor group 2 is composed of sensors having different characteristics such as measurement accuracy. However, the present invention is not limited to this. Is also good. In this case, since the measurement accuracy of the same type of sensor is usually the same, the trajectory calculation can be executed in the same manner as described above only by setting the sensor accuracy coefficient sk to be constant or 1 (substantially no weighting is applied to the sensor accuracy coefficient sk). ). Also in this case, the overall error of the coefficient of the trajectory equation can be reduced as much as possible by multiplying the observation equation by the accuracy coefficient τk relating to the time having a larger value as the time is closer to the current time and using the past data. The effect that can be obtained is obtained.

In the first embodiment, the sensor group 2
Is composed of a plurality of sensors, but may be composed of only a single sensor. In this case, the trajectory calculation can be performed in the same manner as described above only by setting the sensor accuracy coefficient sk to be constant or 1 (substantially no weighting is applied to the sensor accuracy coefficient sk). Also in this case, the overall error of the coefficient of the trajectory equation can be reduced as much as possible by multiplying the observation equation by the accuracy coefficient τk relating to the time having a larger value as the time is closer to the current time and using the past data. The effect that can be obtained is obtained.

Furthermore, in the first embodiment, the case has been described where the trajectory of the horizontal position of the moving body in the orthogonal coordinate system (X coordinate, Y coordinate) is obtained. If it is from the type, the data of the moving object obtained from multiple sensors with different characteristics are decomposed into the polar coordinate distance and azimuth at the horizontal position, and the measurement accuracy of the sensor is considered for each observation formula By multiplying by the sensor accuracy coefficient sk, and further by multiplying the data closer to the current time by the accuracy coefficient τk relating to the time having a larger value, the data may be comprehensively processed to calculate the trajectory. Can be obtained.

Further, the moving object monitoring apparatus according to the first embodiment may be configured such that altitude information (for example,
Mode C beacon code). Therefore, based on this altitude information, the measurement accuracy of the sensor was also considered in the observation formula of the vertical position on the altitude coordinate axis included in the data of the moving object obtained from a plurality of sensors having different characteristics in the same manner as described above for the altitude coordinate axis. Sensor accuracy coefficient s
By multiplying the observation formula by multiplying the observation formula by multiplying the observation formula by multiplying the observation formula by multiplying the observation formula by multiplying the observation formula by multiplying the observation formula by the accuracy coefficient τk having a larger value for the data closer to the current time, and calculating the trajectory, Good tracking can be realized.

Furthermore, in the first embodiment, for the purpose of speeding up the operation, data relating to a large number of moving objects to be monitored are collectively processed and the vector calculation is executed in the trajectory calculation. If the processing capacity is sufficient to perform the above calculation and to process the target number of moving objects, the same effect as described above can be obtained by performing the trajectory calculation by sequential processing. Play.

In the first embodiment, the trajectory calculation is performed using four motion models in each coordinate of the rectangular coordinate system. However, according to the purpose of the calculated moving object monitoring device and the capability of the processing processor, It is also possible to perform simplification such as reducing the number of motion models or using a single motion model. In particular, in the case of sequential processing, such simplification is effective.

Embodiment 2 FIG. 7 is a flowchart showing processing of the moving object monitoring method according to the second embodiment of the present invention. A specific example of the mobile monitoring apparatus that realizes the mobile monitoring method according to the second embodiment has the same configuration as that according to the first embodiment shown in FIG. Only a description will be given, and duplicate description will be omitted. Further, the various modifications described in the first embodiment can be similarly applied to the second embodiment.

Next, the operation will be described. In the first embodiment, the position information on the horizontal position input from each sensor of the sensor group 2 is based on the premise that the error is dispersed around the true position. However, in the second embodiment, Assume a special case.

In the case of a rotating body type sensor, the position information is usually input to the target input device 3 of the moving body monitoring device as polar coordinates. In this case, it is expected that the position information about the moving body to be output is shifted in the azimuth and distance directions as a whole, but the value of the average deviation is unknown and the position of most moving bodies is determined. It can be assumed that this is the sensor that can be observed and has the largest number of data inputs (this sensor is sensor A). On the other hand, it is assumed that there is a sensor whose measurement accuracy is extremely higher than that of the sensor A but the number of data inputs is small, and that the sensor A alone is insufficient for predicting the trajectory (this sensor is referred to as a sensor B).

In the case of the moving object monitoring apparatus provided with such a combination of the sensors A and B, the smoothing prediction processing control unit 6 and the parallel calculation processing unit 7 have already described the first embodiment using only the data from the sensor A. It is assumed that the formula of the trajectory and the value of the coefficient are calculated according to the processing of steps ST301 to ST313 of the flowchart shown in FIG. The expression of the latest trajectory for a certain tracking target obtained in this way is G
_ax (t) and _Gay (t).

Next, in step ST401, when data is input from the sensor B and the data is associated with the above-mentioned tracking target, the position information obtaining time of the input data is set to t.
_n , and the coordinates of the horizontal position in the rectangular coordinate system are X _btn and Y _btn . Next, in step ST402, the rectangular coordinates X
_btn , Y _btn are converted into polar coordinates R _btn , AZ _btn , and in step ST403, the position information obtaining time t _n of the input data of the sensor B is calculated using the above-described latest trajectory equation G _ax (t), G _{ay of the} sensor A. (T), and the calculation position X _atn = G
_ax (tn), obtaining a _{Y atn = G ay (t n} ). Further, in step ST404, the calculated positions X _atn and Y _atn of the orthogonal coordinate system thus obtained are converted into polar coordinates,
R _atn and AZ _atn are obtained.

Then, in step ST405, the smoothing prediction processing control device 6 determines whether the position obtained from the sensor B in the polar coordinate system and the position obtained from the trajectory formula based on the past data of the sensor A Difference in distance direction and azimuth direction,
_{DR n = R btn -R atn,} DAZn = AZ btn -AZ
get _atn . Next, in step ST406, it is determined whether or not the calculation of the difference value is the first. If it is the first time, in step ST407, the trajectory of the trajectory is determined based on the position obtained from the sensor B and the past data of the sensor A. The difference DR in the distance direction and the azimuth direction from the position obtained from the equation
_n and DAZ _n are registered as DR and DAZ, respectively.
On the other hand, if the first non, in step ST 408, the conventional differential DR already registered, DAZ and latest differential DR _n, the DAZ _n performs average processing with the weight according to the following equation to correct the difference.

DR = DR × r + DR _n × (1-r) (20) DAZ = DAZ × r + DAZ _n × (1-r) (21)

Here, r is a weight that is predetermined in the mobile monitoring apparatus and has a value greater than 0 and less than 1 when averaging the latest difference and the past difference. That is, the sensor B
Each time data is obtained, the difference in the distance direction and the azimuth direction between the position obtained from the sensor B in the polar coordinate system and the position obtained from the trajectory formula based on the past data of the sensor A is calculated. For this and the past differences DR and DAZ, (2
The average value of the difference is obtained by performing weighted averaging processing according to the expressions (0) and (21), and this is estimated as a deviation.

[0111] Next, in step ST 409, the data relating to another mobile formulas trajectory of a moving body determined based from the sensor A, the coordinates X of another mobile body at a time t _{m ctm} = G _{cx (t m), Y ctm =} G cy (t
_m ) is obtained. In step ST410,
These coordinates _Xctm , _Yctm are _converted to polar coordinates _Rctm , _AZctm.
In step _ST411 , the differences DR and DAZ obtained in step _ST408 are converted to polar coordinates _Rctm , A
Correction is performed in addition to _Zctm . Then, step ST41
2, the obtained new polar coordinates R ′ _ctm (= R _ctm
+ DR), AZ ' _ctm (= AZ _ctm + DAZ) are converted to rectangular coordinates X' _ctm , Y ' _ctm , which are output to the mobile information processing device 8 as corrected horizontal positions. Note that the difference does not necessarily need to match the coordinate axis used for the calculation.For example, even if the formula of the trajectory is obtained on the orthogonal coordinate axis, if a deviation is expected in the azimuth direction, the orthogonality is used only for the difference calculation. It is also possible to convert the coordinates to polar coordinates and obtain them.

As described above, according to the second embodiment, it is possible to estimate in real time the value of the displacement of the position where the size unique to the sensor A is unknown, and in consideration of this, the formula of the trajectory of another moving object is calculated. Is also corrected, and an effect of realizing accurate tracking can be obtained.

Embodiment 3 FIG. 8 is a flowchart showing processing of the moving object monitoring method according to the third embodiment of the present invention. A specific example of the mobile monitoring apparatus that implements the mobile monitoring method according to the third embodiment has the same configuration as that according to the first embodiment shown in FIG. Only a description will be given, and duplicate description will be omitted. Further, the various modifications described in the first embodiment can be similarly applied to the third embodiment.

Next, the operation will be described. First, in step ST501, the moving object monitoring apparatus uses the smoothing prediction processing control device 6 and the parallel calculation processing device 7 to perform an effective operation based on the data output from the sensor group 2 as in the first embodiment. The trajectory calculation is executed for the horizontal position of the tracking target. As described above, the coefficients of the four trajectory formulas (movement models) are determined by the least square method for each coordinate of the horizontal position of each tracking target, and which trajectory formula is determined at each coordinate of the horizontal position of each tracking target. Whether or not to adopt is determined by evaluating the past fluctuation of the estimated speed. Then, from the coefficients of the adopted trajectory formula, the position of each tracking target,
Velocity v _m and the acceleration a _m is obtained. For example, if each as an expression of the trajectory of the horizontal position of the orthogonal coordinate system of the moving body described above (2) and (8) is selected, the speed v _m and the acceleration a _m is expressed by the following equation.

[0115] ^{_{v m = (v x2 2 +}} v y2 2) 1/2 (22) a m = 2 × (a x2 2 + a y2 2) 1/2 (23)

[0116] However, the acceleration a _m of the tracking target is estimated that performs uniform motion in the X and Y coordinates are zero.

Next, in step ST502, GP
When the target input device 3 of the moving object monitoring device receives data on the velocity v _S and acceleration a _S measured by the moving object itself using techniques such as _S and the data on the horizontal position via the sensors of the sensor group 2, velocity v _s and acceleration a
_{If s} corresponds to the coordinate system of the mobile monitoring device,
In step ST503, the above-mentioned step ST501 is performed.
Velocity v _m and the acceleration a _m and the velocity v _s and the acceleration a _s velocity measured by the mobile itself performs average processing with a weight according to the reliability of the data to and v _s and the acceleration a _s THAT obtained Is corrected. That is, assuming that the weights of the weighted averaging process of the velocity and the acceleration are α _v and α _a , respectively, the corrected velocity v and acceleration a are

[0118] _{v = α v × v s +} (1 one _{α v) × v m (24} ) a = α a × a s + (1-α a) × a m (25)

Is calculated. This makes it possible to obtain the speed and acceleration of each moving object with higher accuracy.

In the case where the speed and acceleration with respect to the air measured by the moving object itself can be obtained via the sensor without corresponding to the coordinate system of the moving object monitoring apparatus, the first embodiment is used.
By calculating the difference between the speed and the acceleration obtained according to the moving body monitoring method according to the above, the direction and speed of the wind can be estimated.

As described above, according to the third embodiment, accurate tracking can be realized, and the effect of monitoring the approach of a moving object and predicting the target arrival time can be obtained early and accurately.

[0122]

As is evident from the foregoing description, according to the present invention, when the type of sensor that outputs the position information in a very seat <br/> mark is common coordinate for finding the trajectory about the horizontal position of the movable body The observation formula for the X coordinate of the system is multiplied by a weight inversely proportional to the distance in the Y direction of the local coordinate system of the sensor, and the observation formula for the Y coordinate is weighted inversely proportional to the distance in the X direction of the local coordinate system of the sensor. Since the multiplication is performed, data from the sensor of the rotating body type and data from other sensors having different characteristics are comprehensively processed to achieve an accurate tracking.

According to the present invention, in the case of a sensor that outputs position information in polar coordinates, the time at which the data was measured is estimated from the azimuth included in the position information. Even if the sensor does not have, the time when the data was measured can be obtained,
There is an effect that accurate tracking can be realized by comprehensively processing data from various sensors including these data.

According to the present invention, the motion adopted as the equation of the trajectory of the moving object is evaluated by evaluating the past speed fluctuation of the moving object based on the equation of the trajectory of the moving object obtained by the calculation of the least square method. Since the configuration is such that the model is determined, there is an effect that a motion model can be selected in consideration of the past acceleration motion and accurate tracking can be realized.

According to the present invention, since the trajectory formula for evaluating the fluctuation of the speed of the moving object is a motion model represented by a quadratic function with respect to time, the past acceleration motion is considered. This makes it possible to select a motion model and achieve accurate tracking.

According to the present invention, as a result of the evaluation of the past speed fluctuation of the moving body, if the moving body continues to move at a constant speed within a certain period in the past, the trajectory is expressed by a linear function with respect to time. If you select the motion model represented and if you keep changing the speed, select the motion model represented by the quadratic function with respect to time as the trajectory formula, and if you start changing the speed from constant velocity motion, Select a motion model represented by a quadratic function with respect to the most recent time as an expression, or, if moving from a state in which the speed is continuously changed to constant velocity motion, use a linear function with a primary function as a trajectory expression. Since the configuration is such that a motion model to be selected is selected, a motion model can be selected in consideration of past acceleration motion, and there is an effect that accurate tracking can be realized.

According to the present invention, the equation of the trajectory of the moving object is estimated using the least squares method based on the position information from the first sensor that outputs the position information including the error whose size is unknown. The error is estimated by applying the measured time of the position information input from the second sensor, which is more accurate than the first sensor, or the time estimated to be measured to the estimated trajectory formula. With this configuration, there is an effect that the value of the displacement of the position whose magnitude unique to the sensor is unknown can be estimated in real time.

According to the present invention, the equation of the trajectory of the other moving object is estimated using the least square method based on the position information of the other moving object from the first sensor, and the estimated error is calculated. Since it is configured to correct the trajectory formula of the other moving object in consideration of the above, the value of the deviation of the position where the size specific to the sensor is unknown can be estimated in real time, and in consideration of this, the expression of the trajectory of the moving object is Has the effect of being able to realize accurate tracking by correcting the.

According to the present invention, every time position information is obtained from the second sensor, the trajectory of the moving body obtained by using the least squares method based on the position information from the first sensor is used. Since the error is estimated again, the error estimated this time and the error estimated in the past are weighted and averaged, and the value is newly estimated as an error. There is the effect that the value of the unknown position shift can be estimated in real time, and the formula of the trajectory of the moving object can be corrected in consideration of this to realize accurate tracking.

According to the present invention, when displaying the positions of a plurality of moving objects on the display device, the positions are obtained by applying the display times to the equations of the trajectories estimated for the plurality of moving objects. With this configuration, the relative position between the moving objects can be accurately displayed, and the position of the moving object can be accurately superimposed on the analog data.

According to the present invention, a fixed amount of previously obtained position information is included for a plurality of moving objects that need to estimate the trajectory equation using the least squares method, and the same A trajectory calculation table to which an area of a size is allocated is provided, and dummy data is written in cells of the trajectory calculation table in which no position information exists, so that parallel calculation processing can be effectively used to perform trajectory calculation. There is an effect that can be executed at high speed.

According to the present invention, each cell of the trajectory calculation table includes the weight relating to the sensor accuracy of the sensor that obtained the position information, and the weight is zero in the cell in which the dummy data is written. With the configuration, there is an effect that the parallel calculation processing can be effectively used and the trajectory calculation can be executed at high speed.

According to the present invention, while data including position information input from a plurality of sensors is temporarily stored,
A tracking table in which dummy data is written is provided in each cell where no data exists, and each time a trajectory formula for a plurality of moving objects is estimated, the (trajectory formula should be estimated once stored in the tracking table). The oldest data of the same amount as the latest data amount of (number of moving objects) × (maximum number of data accumulated for each moving object since the previous trajectory estimation) is deleted from the trajectory calculation table. Since the configuration is such that the latest data is transferred from the tracking table to the trajectory calculation table, there is an effect that the parallel calculation processing can be effectively used and the trajectory calculation can be executed at high speed.

According to the present invention, when the speed and acceleration measured by the moving object are input, the speed and acceleration are obtained from the trajectory formula estimated for the moving object, and the speed and acceleration measured by the moving object itself are obtained. The weighting and averaging are performed to correct the speed and acceleration measured by the moving body itself, so that accurate tracking can be realized, and the approaching monitoring of the moving body and the target arrival time prediction can be performed early. There is an effect that can be performed accurately.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of a mobile monitoring apparatus that realizes a mobile monitoring method according to Embodiment 1 of the present invention;

FIG. 2 is a flowchart showing a process for obtaining input data acquisition time in the moving object monitoring method according to the first embodiment of the present invention.

FIG. 3 is an explanatory diagram showing a structure of a tracking table according to the first embodiment shown in FIG. 1;

FIG. 4 is an explanatory diagram showing an error in the azimuth direction of a rotator-type sensor.

FIG. 5 is a flowchart illustrating a process of obtaining an equation of a trajectory of a moving object in the moving object monitoring method according to the first embodiment.

FIG. 6 is an explanatory diagram showing the structure of a trajectory calculation table according to the first embodiment shown in FIG. 1;

FIG. 7 is a flowchart illustrating a process of estimating a sensor shift and correcting a formula of a trajectory of a moving object in the moving object monitoring method according to the second embodiment of the present invention.

FIG. 8 is a flowchart showing a process of correcting a speed and an acceleration measured by the moving object itself in the moving object monitoring method according to the third embodiment of the present invention.

FIG. 9 is a block diagram showing a tracking device using a conventional moving object monitoring method.

[Explanation of symbols]

5 tracking table, 9 display device, 71 trajectory calculation table.

──────────────────────────────────────────────────続 き Continuation of the front page (56) References JP-A-3-245081 (JP, A) JP-A-3-248073 (JP, A) JP-A-3-122582 (JP, A) JP-A Sho 54- 124693 (JP, A) JP-A-7-104057 (JP, A) JP-A-1-257286 (JP, A) JP-A-1-92678 (JP, A) JP-A-64-65476 (JP, A) JP-A-64-29787 (JP, A) JP-A-5-59375 (JP, U) (58) Fields investigated (Int. Cl. ⁶ , DB name) G01S 7/00-7/42 G01S 13/00 -13/95 G01C 21/20 G08G 5/00

Claims (13)

(57) [Claims] In a moving object monitoring method for estimating a trajectory of a moving object based on position information of the moving object input from each of a plurality of sensors by using a least squares method, a position is measured in polar coordinates.
In the case of sensors that output location information,
X-coordinate of the common coordinate system for finding the trajectory for the horizontal position
The observation formula for the target is Y in the local coordinate system of the sensor.
Multiply the distance in the direction by a weight inversely proportional to the
The measurement formula is the distance in the X direction of the local coordinate system of the sensor.
A moving object monitoring method, wherein a trajectory of the moving object is estimated by multiplying the weight by an inversely proportional weight. 2. The movement according to claim 1 , wherein in the case of a sensor that outputs position information in polar coordinates, a time at which the position information is measured is estimated from an azimuth included in the position information. Body monitoring method. 3. A motion model to be adopted as an equation of the trajectory of the moving object by evaluating a past change in speed of the moving object based on an equation of the trajectory of the moving object obtained by the calculation of the least square method. 3. The moving object monitoring method according to claim 1 or 2, wherein the determination is made. 4. The moving body monitoring method according to claim 3, wherein the trajectory formula for evaluating the fluctuation of the speed of the moving body is a motion model represented by a quadratic function with respect to time. 5. As a result of evaluating the past fluctuation of the speed of the moving object, if the moving object continues to move at a constant speed within a certain period in the past, the motion expressed by a linear function with respect to time as a trajectory expression. If the model is selected and the speed is continuously changing, a motion model represented by a quadratic function is selected as the trajectory formula with respect to time as the trajectory formula. If you select a motion model expressed by a quadratic function with respect to time, or move to a constant velocity motion from a state where the speed is continuously changed, a motion model expressed by a linear function with a recent time as a trajectory formula The mobile object monitoring method according to claim 3 or 4, wherein the method is selected. 6. The first sensor when there is a first sensor that outputs position information including an error whose size is unknown and a second sensor that is more accurate than the first sensor. From the position information from, the formula of the trajectory of the moving object was estimated using the least squares method, and the time at which the position information input from the second sensor was measured or the time at which it was estimated to be measured was estimated. The moving body monitoring method according to any one of claims 1 to 5 , wherein the error is estimated by applying the error to a trajectory formula. 7. A method of estimating the trajectory of the other moving object using a least squares method based on the position information of the other moving object from the first sensor, and taking the estimated error into account. 7. The moving object monitoring method according to claim 6, wherein the equation of the trajectory of the moving object is corrected. 8. Every time position information is obtained from the second sensor, an error is re-estimated from the trajectory formula of the moving body obtained by using the least squares method based on the position information from the first sensor. ,
This estimated the error and mobile object monitoring method according to claim 6 or claim 7, wherein the past and estimated error by averaging with a weighted and estimates the value as a new error . 9. When displaying the positions of a plurality of moving objects on a display device, the positions are obtained by applying display times to equations of trajectories estimated respectively for the plurality of moving objects. The method for monitoring a moving object according to any one of claims 1 to 8 , wherein 10. A region containing a fixed amount of previously obtained position information for a plurality of moving objects for which it is necessary to estimate the trajectory formula using the least squares method, and having the same size for each moving object. provided the trajectory calculation table assigned, moves to a cell of the trajectory calculation table location information is not present claims 1, wherein the writing dummy data according to any one of claims 9 Body monitoring method. 11. Each cell of the trajectory calculation table includes a weight relating to sensor accuracy of a sensor that has obtained position information, and the weight is zero in a cell in which dummy data is written. The method for monitoring a moving object according to claim 10 . 12. A method for temporarily storing data including position information input from a plurality of sensors, providing a tracking table in which dummy data is written in each cell where no data exists, and calculating a trajectory formula for a plurality of moving objects. (Estimation of trajectory formula, number of moving objects to be estimated once) × (Maximum number of data accumulated for each moving object since last trajectory estimation) 11. The method according to claim 10 , further comprising: deleting oldest data having the same amount as the latest data amount from the trajectory calculation table, and transferring the latest data from the tracking table to the trajectory calculation table. Alternatively, the moving object monitoring method according to claim 11 . 13. When the speed and acceleration measured by the moving object are input, the speed and acceleration are obtained from the equation of the trajectory estimated for the moving object, and the speed and acceleration measured by the moving object and the speed and acceleration are calculated. each moving object monitoring according to any one of claims 1 to 12, characterized by correcting the speed and the acceleration measured by the mobile itself by averaging with a weighted Method.