JP3246822B2 - Mobile tracking device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a tracking apparatus for a moving object, and more particularly to a method for determining a moving object model in real time using information from a sensor when a moving model of the moving object is unknown on the observation side. In addition, the present invention relates to a mobile object tracking device that notifies an operator of information on the determined mobile object.

[0002]

2. Description of the Related Art In multi-sensor tracking in which a moving object is tracked by using a plurality of sensors, a device that performs correlation processing of a target and displays the result on a CRT or a head-up display is widely used. Above all, various proposals have been made for devices relating to position determination. For example, in a “position determining device” disclosed in Japanese Patent Application Laid-Open No. 1-274011, the scan time of a one-dimensional array sensor is changed according to the speed of a moving body, and a position is calculated based on an angle with respect to the moving body. I have.

In a "target tracking display device" disclosed in Japanese Patent Application Laid-Open No. 62-12883, maximum likelihood estimation processing is performed using azimuth information such as sound waves and radio waves generated from a moving object, and the position of the target is determined. Is calculated. In addition, this technology uses a passive sensor that does not transmit any signal from the sensor to the moving body, and provides a detection forecasting device that calculates the detection capability of the passive sensor, so that the presence area of the target that changes every moment is It is a system that displays time series and displays the target exercise status.

Japanese Patent Application Laid-Open Nos. 61-184377 and 61-195381 disclose a "target tracking filter". In the former, state estimation means for independently calculating the target position, velocity, acceleration, and estimation time error from the target position observation value are arranged in parallel, and the estimation time errors calculated by the state estimation means are compared, A technique for selecting a target position, velocity, and acceleration by a state estimating unit that minimizes an estimation error is disclosed. In the latter, an estimated value estimated by a similar state amount estimating unit is applied to a least square method, and the square of the error is calculated. There is disclosed a technique for selecting an output value of a state estimating unit that outputs an estimated value that minimizes the sum of.

[0005]

SUMMARY OF THE INVENTION Incidentally, Japanese Patent Application Laid-Open No. Hei.
Japanese Patent Publication No. 274011 discloses a method of calculating its own three-dimensional position based on an angle, using a one-dimensional array sensor, and changing the scan time of the one-dimensional array sensor so as to eliminate a detection error in a traveling direction. However, the calculation is performed based on the coordinates of the one-dimensional array sensor that has detected the target object. Is also not considered.

[0006] Japanese Unexamined Patent Publication No. 62-12883 discloses a position detection using a passive sensor.
When displaying the detected position, the error fluctuation is also displayed, but in the display of the error fluctuation, only the display of the passive information is touched, and when a heterogeneous sensor such as an active sensor is combined, No consideration was given. In addition, the publication discloses a system for calculating an estimation error of a moving object, but the error changes in time series every time information on the moving object is obtained from a sensor. As described above, no particular consideration is given as to whether the range of the error in the estimated position and the estimated speed is displayed on an output device such as a CRT, and the operator is notified of the process of changing the error.

[0007] JP-A-61-184477 and JP-A-6-1984
In Japanese Patent Application Laid-Open No. 1-195381, a tracking radar is disclosed as a sensor, and a position, a velocity, and an acceleration are disclosed as state quantities input to a filter, and other sensors such as an azimuth, a distance, or an azimuth and a distance are integrated. No consideration is given to the sensor that detects the detected object, and no acceleration is required as the state quantity input to the filter.
There is no consideration about processing only with position and speed.

In general, when the result of determination of a moving object is ambiguous, the operator must make a final decision quickly while looking at a large amount of information obtained from sensors.
However, the operator arbitrarily calls information necessary for judging the state of the moving body, and the operator makes a judgment based on the information including errors, and displays estimated information based on the judgment result. This is not disclosed in any of the above-described conventional examples, and there is no consideration regarding this point.

The present invention has been made in view of such a situation of the prior art, and has as its object to detect only an azimuth and a distance as state quantities and to filter the detected data in real time from filter output information obtained by performing a filtering process. It is an object of the present invention to provide a moving object tracking device capable of estimating a target motion model. Still another object is to provide a mobile device in which the operator can arbitrarily select necessary information from display information including error information necessary for determining the state of the moving object, and can perform tracking in consideration of the empirical judgment of the operator. It is to provide a body tracking device.

[0010]

In order to achieve the object of the present invention, in a moving object tracking apparatus provided with a detecting means, an information processing means, and an output means, the detecting means comprises:
It is constituted by one or a plurality of sensors, and outputs sensor data obtained by observing the momentary movement of the moving body by each sensor, and the information processing means, at each time when the sensor data is given, A filtering process is performed on the sensor data with a filter using a plurality of different motion models, and a discrimination process of the motion model of the moving object is performed based on the filtering process result, and the most suitable one motion model is selected. Estimate the position and velocity of the moving object from the processing result of the filter corresponding to the motion model, use the estimated position and velocity as the initial estimated value of the filtering processing at the next time, and The speed is output to the output device, and the output device displays the estimated position and speed. In addition, input means may be provided in addition to this, in which case,
The input means allows the operator to determine the state of the moving body based on the information displayed on the output device and select predetermined items, and is used when the operator needs to make a determination.

[0011]

In the above means, each sensor of the detecting means outputs observation direction data and observation distance data of the moving body to the information processing means. In the information processing means, a number indicating the type of sensor and data corresponding to the input parameter of the sensor are added as pre-processing to be filtered, and filtering processing is performed by a filter using a plurality of different motion models, and estimation of a target is performed. Calculate the covariance of the position, estimated speed, and estimation error, and further use the variation process of the covariance of the estimation error as a criterion for determining the motion model of the moving object. Feedback to all filters as the initial estimated value of the filtering process of the time, and the prediction using the name of the motion model determined as the value of the covariance of these position, velocity and estimation error, and the determined motion model Information such as a position is displayed on a screen or the like using an output unit. Further, when another exercise model is selected by the input means, the specified estimation result is displayed on the output device. Thus, by preparing a plurality of filters having different motion models, it is possible to sufficiently cope with a complicated non-linear motion model of a moving body.

Further, even if the sensors are heterogeneous plural sensors including passive data and active data, the passive information and the active information can be processed by the same filter by performing the pre-processing for filtering, so that more accurate tracking is performed. It becomes possible. Furthermore, by displaying the value of the covariance error in real time each time it is output from the filter, it is possible to display the variation process of the error to the operator. In addition, the motion model identification name and the corresponding covariance error value are displayed on the display screen, and by adding an input device, the operator can display on the display screen even if the determination result of the moving object is ambiguous. It is possible to determine the motion of the moving body while looking at the value of the determined covariance error, and to output estimation information based on the determined motion model to the display device.

[0013]

DESCRIPTION OF THE PREFERRED EMBODIMENTS One embodiment of the present invention will be described in detail with reference to the accompanying drawings. FIG. 2 is a diagram showing a schematic configuration of a moving object tracking device (hereinafter, referred to as a “tracking system”) to which the present invention is applied. This system is
A plurality of sensors 2 receive information such as the azimuth and distance between the moving object and the observation point with respect to the moving object 1 that moves in space, and the moving object 1 whose movement model is unknown on the observation side. On the other hand, by processing the sensor data in the information processing device 9, the position, speed, and the like of the moving body 1 are estimated.

Reference numeral 1 denotes a moving body, and the moving body 1
It always predicts the position and speed of the target point 100 moving straight toward the target point 100 or moving, and following the target point 100 moving based on the predicted value, defines the direction in which the vehicle is going. Exercise while following the exercise model to do. In addition, the motion model of the moving object is unknown on the observation side.

Reference numeral 2 denotes a sensor group as a detecting means including a plurality of sensors 20. The output information from each of the sensors 20 is a distance, an azimuth, and the like to the moving object 1 observed by the sensors 20. Reference numeral 9 denotes an information processing device as information processing means. The information processing device 9 processes data output from the sensor group 2 and performs a process of estimating the position, speed, and the like of the moving body 1. Reference numeral 7 denotes an input device as input means, and reference numeral 8 denotes an output device as output means. The output device 8 is an information processing device 9 according to a condition input by the operator 10 using the mobile tracking system. The state of the moving body 1 and the like, which are the outputs of the above, are displayed. This allows the operator 10 to grasp the status of the mobile unit 1 from the displayed output information. This is the schematic configuration of the target tracking system.

Next, the tracking system will be described by dividing it into individual blocks. First, details of the information processing apparatus 9 will be described with reference to the block diagram of FIG. The sensor group 2 uses a plurality of sensors 20-1 to 20-n (note that the sensors are collectively indicated by the reference numeral 20 as described above), and the distance, orientation, and the like of the moving body 1 whose motion model is unknown. And the individual sensors 20 output detection data corresponding to the characteristics of the sensors to the sensor data processing device 3.

In the sensor data processing device 3, the data output from the plurality of sensors 20 are: A sensor number indicating which sensor 20 has output is given.

2. When data relating to a plurality of moving bodies is input by one sensor 20, a number corresponding to the moving body 1 is assigned.

3. A reception time and the like at which the data was received are given to each data.

4. In addition, a reference value is set in advance, and data that does not fall within the range of the reference value is automatically deleted as erroneous information.

And the like. In this embodiment, a plurality of sensors 20 are used as the sensor group 2. However, even when the number of sensors 20 is one, that is, a single sensor configuration, tracking can be performed by the configuration shown in FIG. It is. Therefore, the detecting means is a single sensor 20.
It can also be constituted only by.

When these series of processes are completed, the sensor data processing device 3 outputs data such as reception distance and reception direction from the erroneous information and the like to the filter processing device 4. The filter processing device 4 includes a plurality of filters 40-1 to 40-m, and each of the filters 40 (hereinafter, the filters are generally denoted by reference numeral 40) assumes a motion model of the moving object 1 in advance. This is for estimating the position and speed of the moving object 1 from the sensor data using the motion model. For example, the moving object 1 is moving straight toward the target point 100 or the moving target point 100 is moving. Is a filter that assumes that the position and speed of the vehicle are always predicted, and the direction in which the vehicle is moving is determined while following the predicted values.

Here, the filtering process is performed on the data from the sensor data processing device 3 using all the filters 40. The data output from each filter 40 is the estimated position and estimated speed at the current time, and the filter 40 can also estimate the error covariance every time data is input like a Kalman filter. In the case of a possible filter, a diagonal term component indicating the magnitude of the estimation error covariance is also output. In this case, it is possible to know the dispersion of the observation distance and the dispersion of the observation azimuth of the mobile unit 1 by analyzing the diagonal components of the error covariance matrix in the target estimation processing device 5 described later. In addition, each filter 4
The initial estimated position and the initial estimated speed at 0 are given from the target estimation processing device 5 as described later.

Each filter 40 performs a filtering process each time data such as a distance and an azimuth is output from the sensor data processing device 3 based on initial estimated values such as an initial estimated position and a process estimated acceleration. Update position and velocity estimates for. The estimation result of the position and speed of the moving body 1 calculated in the filter processing device 4 is sequentially output to the target estimation processing device 5.

The target estimation processing device 5 comprises a first CPU (hereinafter referred to as “CPU (1)”) 51, a state calculation processor 53, and a state processing data table 52. The CPU (1) 51 manages input / output of data flowing in the target estimation processing device 5, and controls data transmission / reception timing and the like. Next, in the state processing data table 52, the data output from the filter processing device 4 is classified for each filter 40 and stored in the table. Here, the format of the state processing data table 52 (hereinafter, the state data table is generally indicated by reference numeral 52) is shown in FIG. 3, and details of the data table 52 will be described with reference to FIG.

The state processing data table 520-k stores information corresponding to the filter 40-k in the filter processing device 4. Here, k = 1 to m, and m is the number of filters (the same applies hereinafter). State processing data table 520
-K is the exercise model identification name 54-k, reception time 531
-K, estimated position 532-k, estimated speed 533-k, predicted position 534-k, covariance matrix 535-k, estimated position error 536-k, estimated speed error 537-k, estimated speed vector 538-k, and It is composed of each item of variance error 539-k. Here, the data calculated from the filter 40 are the reception time 531-k, the estimated position 532-k, the estimated speed 533-k, the predicted position 534-k, and a filter that outputs up to the estimated error covariance matrix. At 40-k, the value 535-k of the diagonal component of the estimated error covariance matrix is stored on the data table 520-k. Predicted position 534
-K is time-series data of the position of the moving object up to a predetermined future predicted by the filter 40-k. Other items are calculated in the state calculation processor 53.

Next, the calculation method performed in the state processor 53 will be described in detail. First, in the case of a filter for estimating the error covariance, the diagonal term component 535-k of the covariance matrix is called for each filter 40-k,
A covariance error 539-k between the estimated position and the estimated speed is calculated from the diagonal term component, and the state processing data table 520-k
Estimated position error 536 on k and estimated speed error 53
Write to item 7-k.

Next, a method of determining a motion model of the moving body 1 using these data will be described. Moving object 1
Is a filter that outputs the covariance matrix 535-k of the estimation error like a Kalman filter, and determines the rate of change of the covariance matrix 535-k of this error by the moving body 1 Is used as a criterion for the motion model. This method is specifically executed as follows. First, a state vector indicating the state of the moving object 1 is represented by the following equation.

[0029]

(Equation 1)

Based on the state vector, the state of the moving body 1 is estimated. The covariance matrix 535-k is calculated as in the following equation.

[0031]

(Equation 2)

Here, from the value of the diagonal component 535 of the covariance matrix on the state processing data table 52, the covariance error P
The error 539 is calculated by the state calculation processor 53 according to the following equation.

[0033]

(Equation 3)

The value calculated here is the estimated position error 536 when the sensor 20 can measure the distance between the mobile unit 1 and the sensor installation point and the angle between the mobile unit 1 and the sensor installation point. Represents. Further, the estimated position error 536 when the sensor 20 can measure the angle between the moving body 1 and the sensor installation point is calculated according to the following equation.

[0035]

(Equation 4)

Next, a method of determining the motion model of the moving body 1 will be described. The method of determining the motion model can be roughly divided into two methods. The first method is to compare the current value of Perror output from the plurality of types of filters 40 in the state operation processor 53.
In this method, Perror output from each filter 40 is used.
Is determined as the motion model of the moving object 1, and the data such as the filter estimated position and the estimated speed in the state processing data table 52 are called again, and the input / output processing is performed. The data is output to the device 6 and is also output to each filter 40 of the filter processing device 4 for use as an initial estimated value of the filter processing at the next time point.

The second method is to determine the motion model of the mobile unit 1 by performing not only one observation but also a plurality of observations. In this method, the value of the moving average of Perror is used as a criterion for determining the motion model of the moving object 1. Specifically, the state processing data table 5
2. The item of the covariance error 359 stored on 2 is called to the state calculation processor 53 for n consecutive observation data,
The moving average value is calculated, the motion model using the filter 40 having the minimum moving average value is determined as a correct motion model, and the estimated position 53 in the state processing table 52 is determined.
2 and the data such as the estimated speed 533 are called again and output to the input / output device 6 and output to each filter 40 of the filter processing device 4 for use as an initial estimated value of the filter processing at the next time point.

Further, the procedure of the estimation process will be described with reference to the flowchart of FIG. In this procedure, the end of tracking is determined in processing 70, and each processing from processing 70 to processing 77 is executed until tracking is completed.
In the process 71, the data from the sensor 20 is
The processing from the processing 72 to the processing 77 is repeatedly performed for each time input to the. Further, the input data is subjected to a filtering process by all filters 40 corresponding to a plurality of motion models (process 72, process 73). In a process 74, the data output in the process 73 is stored in the state processing data table 5
2 and the estimated position 53 output from each filter 40
2. The estimated speed 533, the diagonal term component 535 of the covariance matrix, and the covariance error 539 calculated by the state calculation processor 53 are stored in the state processing data table 52. In process 75,
In order to determine the value of the covariance error 539 stored on the state processing data table 52 for the motion model of the mobile unit 1, the covariance errors 539-k from all the filters 40 are compared with each other, and the covariance error 539-k is compared. The filter 40 having the minimum value of 539-k is determined as a correct motion model. Therefore,
Even if the motion model is unknown on the observation side, the correct motion model can be easily selected. In the process 76, the reception time 531-k, the estimated position 532-k, and the like of the filter 40-k determined in the process 75 are fed back to all the filters 40 as the initial estimated value of the filter 40 at the next estimated time. In the process 77, the filtering result of the motion model determined in the process 75 is output to the output processing device 8.

Next, the input / output processing unit 6 which performs processing for displaying the result calculated by the target estimation processing unit 5 will be described. As shown in FIG.
Is a control second CPU (hereinafter, referred to as “CPU (2)”) 61, an error range processor 62, an output condition setting unit 63, a coordinate converter 64, and an output data table 6.
5 is comprised.

The input / output device 6 first reads data to be displayed on the output device 8 from the state processing data table 52 in the target estimation processing device 5 to the output data table 65.
The data transfer here does not call the data value directly, but transfers the data address while controlling it by the CPU (2) 61. Therefore, each item of the state processing data table 52, the exercise model identification name 54, the reception time 53
1. Estimated position 532, estimated speed 533, estimated position error 5
36, estimated velocity error 537, covariance matrix 535, velocity vector 538, covariance error 539 and predicted position 53
4, and the exercise model identification name 66, the reception time 651, the estimated position 652,
Estimated speed 653, estimated position error 654, estimated speed error 6
55, the covariance matrix 656, the velocity vector 657, the covariance error 658, and the predicted position 659 are preferably stored in the same location in the memory.

Next, the operator 10 sets output conditions using the input device 7 such as a keyboard or a mouse, inputs the set display conditions to the output condition setting unit 63, and outputs an output data table 65 based on the set conditions. To CPU (2) 6
1 and converted into a form suitable for the output device 8 in the coordinate conversion device 64 and displayed on the output device 8. Also,
When the error range is selected in the output condition, the position error 65
4 and the range of the speed error 655 are calculated by the error range processor 62.

Details of the processing procedure of the input / output processing device 6 will be described with reference to the flowchart of FIG. Here, the tracking result of the moving object and the motion model determination result, which are the contents of the state processing data table 52 already in the target estimation processing device 5, are written in the state output data table 65 in real time. For the k-th filter, the table 650-k includes a filter number 66-k, a reception time 651-k, and an estimated position 65, which are identification names of the model.
2-k, estimated speed 653-k, estimated position error 654
k, estimated speed error 655-k, covariance matrix 656-k,
Estimated velocity vector 657-k, covariance error 658-k,
Predicted position 659-k and motion model determination result 650-
k. Exercise model determination result 650-
As for k, “1” is set for one determined motion model, and “0” is set for other models.

In this procedure, first, a display item setting screen 81 is displayed on the output device 8 as shown in FIG. 6, which is an example of a setting screen. An output item selection area 811 is selected to output information, and a timing at which the selected information is output is set from an output timing setting area 813. The timing here will be described as follows, taking as an example a case where the fluctuation state of the error of the moving object 1 is used as a reference.

The position error (P_err) and the speed error (V
_Err) is set to display the error range on the display device when a <P_err <b or c <V_err <d. The operator 10 selects items of the estimated speed and the estimated position from the output items 811 by using a device such as a mouse.

2. Next, a necessary item is selected from the timing items 813. Here, when the item of “always” is selected, the information is constantly displayed on the screen.

3. As a timing reference, the values of a and b are input using a device such as a keyboard in order to output the position error (P_err) on the screen when the position error (P_err) falls within the range from a to b. Similarly, the values of c and d are input.

The procedure from 4.1 to 3 is repeated, and the operator 10 customizes. Thus, it is possible to indicate to the operator 10 that the analysis accuracy has been improved (display of an error will be described later), and to provide guidance display when the motion model of the moving body 1 changes. Becomes Also, when the item of the judgment motion model is selected, the number of ranks to be output from the motion models judged by the target estimation processing device 5 in the order of the estimated point rank, that is, the covariance error Perror in descending order, is determined. Select from the setting area 812.

As described above, the operator 10 operates the mobile unit 1
An output item and an output timing are arbitrarily selected from the analysis results of the above, and when all the operation of setting the conditions are completed, the operation completion symbol 814 is designated. Operation complete symbol 81
When 4 is specified, all data is output to the output condition setting unit 63 as output option information. In the output condition setting unit 63, the information of the item selected by the input output option information is read from the output data table 65 by the CPU.
(2) The moving object information is output to the coordinate converter 64 via 61. Further, when the moving body information includes a position error and a speed error, the error range processor 62 calculates what distribution the error range has. The calculation method is to change the direction of the velocity vector from the value of the estimation error
When the axis and the orthogonal coordinates are the x-axis, the ellipse is as shown in Expression 5.

[0049]

(Equation 5)

The value calculated by Expression 5 indicates the range of the covariance of the estimated position, and the range of the covariance of the speed is calculated by the same procedure. The result of the error range processing unit 62 is output to the coordinate converter 64.

Next, the coordinate converter 64 performs pre-processing for displaying in accordance with the format of the output device. For example, when the output device 8 is a CRT, the coordinate converter 64 determines which shape, color, and line type to output the moving object information (estimated position, estimated speed, etc.) based on preset information. Processing is performed to determine what color is output at the position. Further, it is possible for the operator to use the input device 7 to reset the shape, color, and the like of the estimated position while the process is being performed. Here, an example of the output screen is shown in FIG. The output screen shown in the figure is roughly classified into a graphic area 91 for displaying the calculation results as a graph and the like, and a character area 82. The character area 82 includes an estimated position 821, an estimated speed 822, an exercise model 823, Switch 82 for switching to information or model determination point 824 and item setting screen
5 is displayed.

In the graphic area 91, the route 91 of the moving object using the filter 40 in which the value of the motion model determination result 650-k of the output data table 65 is "1".
1, an estimated position 912, an estimated position error 913, a speed vector 914, an estimated speed error 915, and a predicted route 916 are shown. Here, the predicted route 916 is displayed as a sequence of points at the predicted position 659 for the determined filter.

In the character area 82, an estimated position 821 of the moving object 1 is displayed in latitude and longitude, and an estimated speed 822 is displayed. Also, the currently displayed exercise model identification name 66 is displayed as an exercise model 823, and a covariance error value Perror 824 is also displayed. The switch 825 for switching the display state allows the display state to be displayed on a screen as shown in FIG. 7 or to be switched to the display item setting screen 81 in FIG. Here, in the display of the covariance error Perror 824, several kinds of identification names 66-1 of the motion models are displayed in descending order of the points, that is, the covariance error Perror 824 is small, and the operator 10 arbitrarily switches using the input device. be able to. When switched, CPU
(2) The information of the specified motion model and the predicted route of each filter 40 are displayed from the output data table 65 via 61, the value of the determination result 650-k is set to "1", and the value is output from the other filters 40. The value of the determined result 650-k is rewritten to “0”. In this case, the CPU (2) 6
The designated exercise model identification name 66 is transmitted to the CPU (1) 51 via the CPU 1 and the designated exercise model is replaced by the CPU (1) 51 in place of the exercise model determined by the above-described determination method. As the determined motion model, the position and speed estimated by the motion model may be sent to the filter processing device 4 as initial values of the estimation process at the subsequent time.

Further, the processing procedure of the judgment processing will be described in detail with reference to the flowchart of FIG. The process 70 is a box for determining the end of the tracking, and performs the processes 71 to 79 until the tracking is completed. Processing 7
1 repeats processing 72 to processing 78 each time information such as distance and direction is obtained by each sensor 20-n. Processing 7
2 is a plurality of types of filters 40-m and each sensor 20-n
The information such as the distance and the azimuth output from is input and the filtering processing (processing 73) is performed for all the filters 40-
Execute for k. In processing 74, all filters 4
The filtering results such as the estimated position (532, 652) and the estimated speed (533, 653) output from 0-k are stored in the state processing data table 52 and the output data table 6.
5 is stored. Further, the state calculation processor 53 calculates the covariance error (539, 658) and the like from the estimated position (532, 652), the estimated speed (533, 653), the diagonal term component (535, 656) of the covariance matrix, and the like. Are stored in the state processing data table 52 and the output data table 65.

In step 78, the value 658-k of the covariance error Perror calculated in step 73 is extracted from the output processing data table 65 and displayed on the output device 8. Process 79
Then, the operator 10 performs a process of inputting an arbitrary exercise model number from the exercise models displayed on the output device 8 by using the input device 7, and furthermore, outputs the output data table 6.
The value of the determination result 650-k in 5 is written as "1" in the selected exercise model and "0" in other exercise models. The process 76 includes an estimation position 65 selected in the process 79.
The values of 2-k and the estimated speed 653-k are fed back as initial estimated values of the filtering process (process 73) at the next time. In the process 77, the position and speed error ranges are calculated from the error range processor 62, the position error 652-k, and the estimated speed 653-k based on the estimated value selected in the process 79, and the estimated position 652-k and the estimated speed are calculated. 653-k, the velocity vector 657-k, the covariance error 658-k, and the predicted position 659-k are converted into values suitable for the format of the output device 8 and displayed on the output device. Further, the information to be output can be arbitrarily selected by the operator 10 according to the format output from the output condition setting unit 63.

As described above, according to the present embodiment, by applying the moving body motion model discrimination method to the moving body tracking system, even if the moving body motion model is unknown to the observer, the motion model can be obtained. It is possible to determine in real time. In addition, since the fitness (covariance error value) of the estimated motion model is displayed on the output screen in addition to the information of the information position and the speed, the operator can arbitrarily select the motion model. In addition, it is possible to increase the degree of freedom in estimating the motion model of the moving object, and by displaying the position error range, it is possible to visually display the error characteristics to the operator, thereby improving operability. A high tracking system can be achieved.

[0057]

As is clear from the above description, since the present invention is configured as described above, only the azimuth and the distance are detected as the state quantities, and the filter output information obtained by filtering the detected data is filtered. From this, a target motion model can be estimated in real time. In addition, a moving object tracking device that allows an operator to arbitrarily select necessary information from display information including error information necessary for determining the state of the moving object and that can perform tracking in consideration of the empirical judgment of the operator. Can be provided.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a detailed configuration of a moving object tracking device according to an embodiment of the present invention.

FIG. 2 is a system configuration diagram illustrating a schematic configuration of a moving object tracking device according to an embodiment.

FIG. 3 is a diagram illustrating a configuration of a data table in a target estimation processing device in the tracking device according to the embodiment.

FIG. 4 is a flowchart illustrating a procedure of an estimation process in a target estimation processing device and a filter processing device according to the embodiment.

FIG. 5 is an explanatory diagram showing a data flow in the input / output processing device in the mobile tracking device according to the embodiment.

FIG. 6 is a diagram illustrating an example of a screen displayed for setting output conditions in the mobile object tracking device according to the embodiment;

FIG. 7 is a diagram illustrating an example of a screen that displays a state determination result in the moving object tracking device according to the embodiment;

FIG. 8 is a flowchart illustrating a determination processing procedure when an operator makes a determination in the moving object tracking device according to the embodiment;

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Moving body 2 Sensor group 3 Sensor data processing device 4 Filter processing device 5 Target estimation processing device 6 Input / output processing device 7 Input device 8 Output device 9 Information processing device 10 Operator 20 Sensor 40 Filter 51 CPU (1) 52 State processing data Table 53 State calculation processor 54 Identification name of motion model 61 CPU (2) 62 Error range processor 63 Output condition setting device 64 Coordinate converter 65 Output data table 66 Identification name of motion model 81 Display item setting screen 82 Character area 91 Graphic area 100 target points

Continued on the front page (72) Inventor Toshiro Sasaki 1099 Ozenji Temple, Aso-ku, Kawasaki City, Kanagawa Prefecture Inside Hitachi, Ltd.System Development Laboratory (72) Inventor Fuminobu Furumura 1099 Ozenji Temple, Aso-ku, Kawasaki-shi, Kanagawa Hitachi, Ltd. Within the System Development Laboratory (72) Inventor Hiroshi Takahashi 216 Totsuka-cho, Totsuka-ku, Yokohama-shi, Kanagawa Prefecture Within the Information and Communication Division, Hitachi, Ltd. (72) Inventor Shoji Yoshi ▼ 216 Totsukacho, Totsuka-ku, Yokohama-shi, Kanagawa Prefecture Address Hitachi Information Systems, Ltd.

Claims (5)

(57) [Claims] 1. A moving object tracking apparatus for calculating an estimated position and an estimated speed of a moving object and estimating a state of the moving object. Detecting means for detecting the distance and the distance, and outputting the detected data as detection data. At a time when the detection data detected by the detection means is input, filtering processing is performed on the detection data by a filter using a plurality of different motion models. Do
A discriminating process of the motion model of the moving object is performed based on the result of the filtering process, one motion model that is most suitable by the discriminating process is selected, and the position and the moving object of the moving object are determined from the processing result of the filter corresponding to the selected motion model. An information processing means for estimating the velocity, using the estimated position and velocity as an initial estimation value for the filtering processing at the next time, outputting the estimated position and velocity, and outputting from the information processing means Output means for displaying the estimated position and speed. A tracking apparatus for a moving object, comprising: 2. The information processing means outputs an identification name of the selected motion model in addition to the estimated position and speed, and the output means outputs the identification name in addition to the estimated position and speed. The moving object tracking apparatus according to claim 1, wherein an identification name of the selected exercise model is displayed. 3. A moving object tracking device for calculating an estimated position and an estimated speed of a moving object and estimating a state of the moving object. Detecting means for detecting the distance and the distance, and outputting the detected data as detection data; input means for inputting the identification name of the exercise model; Performing a filtering process using a filter using different motion models, calculating a covariance error by the filtering process, outputting a covariance error obtained by each of the filters, and calculating a covariance error of the motion model input by the input unit. The position and speed of the moving object are estimated from the processing result of the filter corresponding to the identification name, and the estimated position and speed are Information processing means used as an initial estimated value of time filtering processing; and the estimated position output from the information processing means;
Output means for displaying a velocity and a covariance error. 4. The tracking apparatus for a mobile object according to claim 1, wherein said detection means comprises one or more sensors. 5. The tracking device according to claim 1, wherein said output means comprises at least one of a display device and a printing device.