JP6910314B2 - Target tracking device - Google Patents

Description

The present invention relates to target tracking devices connected to each other via a network.

A system in which a plurality of target tracking devices are connected by a network and a target is tracked by the plurality of target tracking devices is called a sensor network system. The target tracking device is also called a node.

The target tracking device in the sensor network system is called a partial track integrated target tracking device. The partial track integrated target tracking device generates a partial track based on the observed value obtained from the sensor of the own node, and sends the partial track to another node via the network. In addition, the target tracking device receives the partial wake generated by the other node from the other node. The target tracking device integrates a plurality of partial tracks to generate an integrated track. In addition, the target tracking device fuses the integrated track and the newly generated partial track to generate a new integrated track. By checking the newly generated integrated track, the operator can obtain information on the target tracked by the sensor network.

Patent Document 1 discloses a target tracking device that determines whether to send a partial track to another node via a network based on the upper limit value and the lower limit value of the prediction error of the integrated track. When the prediction error of the integrated track is larger than the upper limit value, the target tracking device described in Patent Document 1 sends a partial track to another node via a network in order to improve the prediction error. The target tracking device described in Patent Document 1 suspends the transmission of a partial track when the prediction error is less than the lower limit value. In the target tracking device described in Patent Document 1, when the prediction error is between the upper limit value and the lower limit value, the improvement amount of the prediction error of the integrated track, which is improved by the node sending the partial track, sets the threshold value. Only when it exceeds, the partial wake is sent to other nodes via the network.

JP-A-2017-58192

However, in the target tracking device described in Patent Document 1, a self-position error, which is an absolute error of the self-position, or a time error occurs between nodes connected by a network. Self-positioning error or time error is also called bias error. In other words, the bias error is the time or position error that occurs between the nodes. When the partial track having the bias error and the integrated track are fused, there is a problem that the tracking accuracy of the newly generated integrated track deteriorates.

The present invention has been made in view of the above, and an object of the present invention is to obtain a target tracking device capable of suppressing deterioration of tracking accuracy of a track.

In order to solve the above-mentioned problems and achieve the object, the target tracking device according to the present invention includes a sensor for observing the target, a tracking processing unit for generating a partial track of the target using the observed value of the sensor, and a plurality of target tracking devices. The integrated track generator that generates the integrated track using the partial track, the communication device that transmits the partial track to the integrated track generator, the difference between the predicted value of the integrated track and the smooth value of the partial track, and the integrated track Using the covariance matrix of the prediction error, the first calculation unit that calculates the first calculation result, which is the spread of the covariance matrix of the prediction error of the integrated track, and the covariance matrix of the smoothing error of the difference and the partial track. The difference between the second calculation unit that calculates the second calculation result, which is the spread of the covariance matrix of the smoothing error of the partial track, and the sum and difference between the first calculation result and the second calculation result. It is characterized by including a bias error determination unit that performs a bias determination for determining whether to send a partial track to a communication device using the above.

According to the present invention, it is possible to suppress the deterioration of the tracking accuracy of the integrated track.

Block diagram showing the configuration of the sensor network system according to the first embodiment Block diagram showing the configuration of the node according to the first embodiment The figure which shows the structural example of the control circuit of Embodiment 1. A flowchart showing a partial wake determination process of the partial wake portion according to the first embodiment. The figure which shows the partial track determination process of the partial track part which concerns on Embodiment 1 in coordinates. Block diagram showing the configuration of the node according to the second embodiment

Hereinafter, the target tracking device according to the embodiment of the present invention will be described in detail with reference to the drawings. The present invention is not limited to this embodiment.

Embodiment 1.
FIG. 1 is a block diagram showing a configuration of a sensor network system according to the first embodiment. The sensor network system 200 includes nodes 100-1 to 100-N and network 2. Nodes 100-1 to 100-N track target 1. Nodes 100-1 to 100-N are each mounted on a moving body such as a ship. Nodes 100-1 to 100-N are referred to as nodes 100 when they are shown without distinction. The network 2 connects nodes 100-1 to nodes 100-N so as to be able to communicate with each other. The network 2 is a wireless network. Nodes 100-1 to Node 100-N share a partial track between Nodes 100-1 to Node 100-N via the network 2.

FIG. 2 is a block diagram showing the configuration of the node 100-1 according to the first embodiment. The configuration of node 100-2 to node 100-N is the same as the configuration of node 100-1. Node 100-1 includes a sensor 10, a tracking processing unit 20, a partial wake unit 30, an integrated wake unit 40, a communication device 50, and a display 60.

The sensor 10 is a radar, an optical camera, or the like, and acquires the observed value of the target 1. The observed value includes information on motion specifications such as position and velocity, and information on the time when the sensor 10 observes the target.

The tracking processing unit 20 acquires the observed value from the sensor 10. The tracking processing unit 20 performs smoothing processing using a plurality of observed values acquired periodically, and calculates the smoothing value. The smooth value is updated after tracking processing such as Kalman filter processing is performed. The tracking processing unit 20 performs prediction processing for predicting the target track using the smooth value, and calculates the predicted value. The predicted value is a value that predicts information about the movement specification of the target. The tracking processing unit 20 generates a partial track using the smooth value and the predicted value. The tracking processing unit 20 outputs the generated partial wake to the partial wake unit 30.

The partial track unit 30 includes a prediction processing unit 31, a track fusion unit 32, a transmission adjustment unit 33, a bias error determination unit 34, a first calculation unit 35, a second calculation unit 36, and a transmission unit 37. And.

The prediction processing unit 31 calculates a predicted value, a prediction error, and a prediction error covariance matrix using the integrated track acquired from the integrated track unit 40. The track fusion unit 32 calculates the smoothing value and the smoothing error by using the predicted value and the prediction error acquired from the prediction processing unit 31 and the partial track acquired from the tracking processing unit 20. The transmission adjustment unit 33 uses the prediction error and the smoothing error to perform the first transmission determination, which is a determination of whether to transmit the partial track to the network 2. The prediction error is calculated using the prediction error covariance matrix. The smoothing error is calculated using the smoothing error covariance matrix. The details of the first transmission determination will be described later.

The result of the first transmission determination performed by the transmission adjustment unit 33 is sent to the bias error determination unit 34 or the transmission unit 37. The bias error determination unit 34 makes a second transmission determination using the result of the first transmission determination and the calculation results of the first calculation unit 35 and the second calculation unit 36. The second transmission determination is to determine whether or not there is an error in the bias of the partial track, and determine whether to transmit the partial track. The second transmission determination is also called a bias determination.

The first calculation unit 35 calculates the spread of the prediction error covariance matrix of the integrated track by using the prediction error covariance matrix of the integrated track and the difference between the predicted value of the integrated track and the smooth value of the partial track. The second calculation unit 36 calculates the spread of the smoothing error covariance matrix of the partial track using the smoothing error covariance matrix of the partial track and the difference between the predicted value of the integrated track and the smoothing value of the partial track. That is, the calculation result is the spread of the prediction error covariance matrix of the integrated track and the spread of the smoothing error covariance matrix. Prediction error of integrated wake The spread of the covariance matrix is also called the first calculation result. The spread of the smoothing error covariance matrix is also called the second calculation result.

The calculation results of the first calculation unit 35 and the second calculation unit 36 are sent to the bias error determination unit 34. The sending unit 37 receives the result of the first sending determination or the result of the second sending determination. The result of the transmission determination indicates either the transmission execution of the partial track or the transmission suspension of the partial track. When the transmission unit 37 receives the determination result of the transmission execution of the partial track, the transmission unit 37 sends the partial track to the communication device 50. When the partial track is sent from the sending unit 37, the smoothing value calculated by the tracking processing unit 20 is initialized. When the transmission unit 37 receives the determination result of the transmission suspension, the transmission unit 37 suspends the transmission of the partial track to the communication device 50.

The integrated wake unit 40 includes an integrated wake generation unit 41 and an integrated wake storage unit 42. When the integrated track generation unit 41 acquires a partial track from the communication device 50, the integrated track generation unit 41 fuses the integrated track stored in the integrated track storage unit 42 with the acquired partial track and updates the integrated track. The integrated track storage unit 42 stores the integrated track updated by the integrated track generation unit 41.

The communication device 50 connects to the network 2 and transmits / receives a partial wake between its own node and another node. Further, the communication device 50 transmits a partial track of the own node or another node to the integrated track unit 40. For example, when the own node is node 100-1, the other nodes are 100-2 to 100-N. The display 60 displays the integrated track.

The tracking processing unit 20, the partial wake unit 30, and the integrated wake unit 40 according to the first embodiment are realized by a processing circuit which is an electronic circuit that performs each processing.

The processing circuit may be dedicated hardware or a control circuit including a memory and a CPU (Central Processing Unit) that executes a program stored in the memory. Here, the memory corresponds to, for example, a non-volatile or volatile semiconductor memory such as a RAM (Random Access Memory), a ROM (Read Only Memory), or a flash memory, a magnetic disk, an optical disk, or the like. When the processing circuit is a control circuit including a CPU, the control circuit is, for example, the control circuit 300 having the configuration shown in FIG.

As shown in FIG. 3, the control circuit 300 includes a processor 300a, which is a CPU, and a memory 300b. When it is realized by the control circuit 300 shown in FIG. 3, it is realized by the processor 300a reading and executing the program corresponding to each process stored in the memory 300b. The memory 300b is also used as a temporary memory in each process performed by the processor 300a.

FIG. 4 is a flowchart showing a partial wake determination process of the partial wake unit 30 according to the first embodiment. The partial track determination process is a process including a first transmission determination and a second transmission determination. Steps S3 to S7 are the first transmission determination processes performed by the transmission adjustment unit 33. Steps S11 to S14 are second transmission determination processes performed by the bias error determination unit 34. The partial track unit 30 executes the first transmission determination process and the second transmission determination process each time the tracking process unit 20 generates a partial track.

The partial track unit 30 acquires the partial track and the integrated track (step S1). The prediction processing unit 31 performs prediction processing using the integrated track and calculates a prediction error (step S2).

The transmission adjustment unit 33 compares the prediction error with the upper limit threshold value of the prediction error (step S3). When the prediction error is equal to or less than the upper limit threshold value of the prediction error (steps S3 and No), the partial track determination process proceeds to step S4. When the prediction error is larger than the upper limit threshold value of the prediction error (steps S3 and Yes), the partial track determination process proceeds to step S8. The upper limit threshold of the prediction error is set according to the transmission request, which is the request for the tracking accuracy of the target. The transmission request is set on the node 100 by the operator. The higher the transmission request, the lower the upper limit of the prediction error.

The transmission adjustment unit 33 compares the prediction error with the lower limit threshold value of the prediction error (step S4). When the prediction error is smaller than the lower limit threshold value of the prediction error (steps S4 and No), the partial track determination process proceeds to step S5. The lower limit threshold of the prediction error is set according to the sending request of the operator.

In step S5, the transmission adjustment unit 33 determines that the prediction error of the integrated track is within the permissible range, and sends a determination result of suspending the transmission of the partial track to the transmission unit 37. The sending unit 37 suspends the sending of the partial wake. Therefore, the integrated wake will not be updated.

When the prediction error is equal to or greater than the lower limit threshold value of the prediction error (step S4, Yes), the partial track determination process proceeds to step S6. The transmission adjusting unit 33 updates the integrated track by fusing the partial track and the integrated track, and calculates the smoothing error using the updated integrated track (step S6). The smoothing error is calculated using the updated integrated track smoothing error covariance matrix.

The transmission adjustment unit 33 calculates an improvement amount of the prediction error of the integrated track from the prediction error and the smoothing error calculated in step S6, and compares it with the threshold value (step S7). When the improvement amount of the prediction error of the integrated track is smaller than the threshold value (steps 7 and No), the partial track determination process proceeds to step S5. The threshold is set according to the operator's transmission request.

When the amount of improvement in the prediction error of the integrated track is equal to or greater than the threshold value (step 7, Yes), the process of step S8 is executed. The first calculation unit 35 and the second calculation unit 36 calculate ΔX, which is the absolute value of the difference between the smooth value of the partial wake and the predicted value of the integrated wake (step S8).

First arithmetic unit 35, prediction error covariance matrix P ^F Integration track _- formed among the ellipse form, a range of a straight line connecting predicted value of the smoothing value and integration track portions track in the orthogonal coordinate space _⁽⁾ It is the ^P _{F (-)} spread size [Delta] P ^F of _(-) is calculated and sent to the bias error decision unit 34 (step S9). Prediction error covariance matrix ^P _F Integration track _(-), and ^P _{F (-)} spread size [Delta] P ^F of _(-) will be described in detail later.

The second calculation unit 36, among the ellipse forming part track of the error covariance matrix P ^{S ₍₊₎} is formed in a range of a straight line connecting predicted value of the smoothing value and integration track portions track in the orthogonal coordinate space It is the calculated a ^P _{S (+)} spread size [Delta] P ^S of the _(+), and sends to the bias error determination unit 34 (step S10). Part track of error covariance matrix ^P _{S (+)} and ^P _{S (+)} will be described in detail later spread of size [Delta] P ^S ₍₊₎ of.

Bias error determination unit 34, ΔP ^F _(-) the sum of the ΔP ^S _(+), and calculates the quotient obtained by dividing the [Delta] X (step S11). [Delta] P ^F _(-) and the quotient of the sum divided by the ΔX and ΔP ^S ₍₊₎ is also referred to as the first quotient. When the first quotient is smaller than 1 (step S11, Yes), the bias error determination unit 34 determines that the partial track has a bias error, and sends a determination result of suspending the transmission of the partial track to the transmission unit 37 (step S11, Yes). Step S12). The sending unit 37 suspends the sending of the partial wake. Therefore, the integrated wake will not be updated.

When the first quotient is 1 or more (steps S11, No), the bias error determination unit 34 determines that there is no bias error in the partial track, and determines that the partial track is sent to the network 2 and the integrated track generation unit 41. The result is sent to the sending unit 37, and the sending unit 37 sends the partial track to the network 2 and the integrated track generation unit 41 (step S13). By fusing the partial track and the integrated track, the error of the integrated track is improved, and the deterioration of the tracking accuracy of the sensor network system 200 is suppressed.

When the partial track is output from the transmission unit 37, the tracking processing unit 20 deletes the partial track and initializes the smoothing value (step S14).

A process of determining whether or not the partial track input to the bias error determination unit 34 has a bias error will be described.

FIG. 5 is a diagram showing the partial wake determination process of the partial wake unit 30 according to the first embodiment in coordinates. The smooth value of the partial wake and the predicted value of the integrated wake are the values of the Cartesian coordinate system. The smooth value of the partial wake and the predicted value of the integrated wake have information such as position and speed. A smoothing error covariance matrix around the smooth value of the partial track, a prediction error covariance matrix around the predicted value of the integrated track, an ellipsoid as the probability distribution, and the ellipsoid of the smooth value and the predicted value. Shows the spread of the error. In FIG. 5, the ellipsoid formed by the broken line shows the smoothing error covariance matrix of the partial wake. The ellipsoid formed by the thick line shows the prediction error covariance matrix of the integrated wake.

When there is a bias error between nodes 100-1 to nodes 100-N, the difference between the smooth value of the partial wake and the predicted value of the integrated wake increases. On the other hand, the magnitude of the spread of the error covariance matrix that spreads around the smooth value and the predicted value is constant without being affected by the bias error. Therefore, the bias error determination unit 34 divides the sum of the spreads of the smoothing error covariance matrix of the partial track and the prediction error covariance matrix of the integrated track by the magnitude of the difference connecting the smoothing value and the predicted value. When the quotient of 1 is smaller than 1, it is determined that the bias error is present.

Partial and integrated wakes include components such as position and speed. The error covariance matrix shows the correlation of errors such as position and velocity. When a request for sending a partial wake is made, the integrated wake is predicted and processed using the time information of the partial wake. The prediction processing of the integrated wake when the difference between the time information of the integrated wake and the time information of the partial wake is Δt is as follows.

^{X F} _{+ in} Eq. (1) is the smooth value of the integrated wake. x ^F ₊ , y ^F ₊ , and z ^F ₊ indicate the target position, and x ^{· F} ₊ , y ^{· F} ₊ , and z ^{· F} ₊ indicate the target speed. ^XF _{− in} equation (2) is the predicted integrated track, and the subscript T indicates transpose. ^P _{F +} is of the formula (3), a smooth error variance matrix of the consolidated wake. P ^F ₋ is a prediction error variance matrix of the integrated track. The bias error determination unit 34 uses the predicted value of the integrated track that has been predicted and processed to obtain the difference ΔX between the smooth value of the partial track and the predicted value of the integrated track.

^{X S} _{+ in} Eq. (4) indicates the smoothness value of the partial track. ΔX in the formula (5) indicates the difference between the smooth value of the partial wake and the predicted value of the integrated wake.

Bias error determination unit 34, the spread of the prediction error covariance matrix of the integrated track connecting the smoothed value and the predicted value magnitude [Delta] P F _- Request _^(). The error covariance matrix is a multidimensional matrix. The error covariance matrix represents the correlation of the variance for each x, y, z component such as position and velocity. The bias error determination unit 34 can extract necessary matrix components by multiplying the error covariance matrix by the observation matrix H. For example, when the error covariance matrix shows the variance between the position of 6 rows and 6 columns and the velocity, when extracting the matrix component related to the position, it is obtained as in Eq. (6).

The error covariance matrix is an ellipsoid that extends around the predicted or smoothed values. By considering the smooth value and the predicted value as two points existing in the three-dimensional space of the Cartesian coordinate system and considering the error covariance matrix as an ellipsoid, the line formed by the smooth value and the predicted value and the ellipsoid can be obtained. The point of contact with the line to be formed can be obtained. Spread [Delta] P F smoothed value and the prediction error covariance matrix of the integrated track which connects the predicted value _^(-) is obtained as equation (7).

Spread [Delta] P S smoothed value and error covariance matrix of connecting it partially track the predicted value _⁽⁺⁾ is, [Delta] P F _- can be obtained in the same manner equation (8) ^_().

If there is a bias error between nodes 100-1 to 100-N, the first is the sum of the spreads of the error covariance matrix divided by the difference between the smoothness of the partial track and the predicted value of the integrated track. The quotient takes a value smaller than 1, and the bias error can be determined by the equation (9).

When the value of the equation (9) is smaller than 1, the bias error determination unit 34 determines that there is a bias error, and the transmission unit 37 suspends the transmission of the partial track. When the first quotient of the equation (9) is 1 or more, the bias error determination unit 34 determines that there is no bias error, and the transmission unit 37 transmits the partial track.

From the above processing, the bias error determination unit 34 determines the bias error and repeatedly executes the transmission control of the partial track.

Although the equation (6) has been described using the information on the position, the bias error determination unit 34 can determine the bias error even when the velocity, acceleration, jerk, etc. are used for ΔX. Further, the bias error determination unit 34 is not limited to any one of speed, acceleration, and jerk as an element of bias error determination, and can determine a combination of a plurality of elements.

As described above, according to the present embodiment, when the tracking processing unit 20 generates a partial track from the observed value having a bias error, the transmission adjusting unit 33 and the bias error determining unit 34 take the partial track. Suspend sending. Therefore, in the present embodiment, since the partial wake including the bias error is not sent to the integrated wake generation unit 41 and the network 2, it is possible to suppress the deterioration of the integrated wake of the nodes 100-1 to 100-N. Can be done.

Further, in the present embodiment, since the partial track transmitted to the network 2 through the communication device 50 is limited, the communication load of the network 2 is improved.

Embodiment 2.
FIG. 6 is a block diagram showing a node configuration according to the second embodiment. The node 100a includes a sensor 10a, a bias error determination unit 34a, a tracking processing unit 20a, a prediction processing unit 31a, a track storage unit 70, and a display 60a. In the present embodiment, the bias error determination unit 34a determines the bias error using the observed value acquired by the sensor 10a and the track stored by the node 100a.

The sensor 10a acquires the observed value from the target and outputs the observed value to the bias error determination unit 34a. The bias error determination unit 34a performs bias error determination processing from the observed value input from the sensor 10a and the track predicted and processed by the prediction processing unit 31a, and determines whether to send the observed value to the tracking processing unit 20a. do. The tracking processing unit 20a performs tracking processing such as Kalman filter processing using the observed values, executes prediction processing and smoothing processing, and generates a track. The generated track is stored and managed in the track storage unit 70. The generated track is displayed on the display 60a as track information.

The operation of the bias error determination performed by the bias error determination unit 34a according to the present embodiment will be described.

Z is an observed value. The observed values are information on motion specifications including the position and velocity of the target 1, and time information. X ₍₋₎ is a predicted value of the track predicted and processed by the tracking processing unit 20a. R is an observation error covariance matrix. Further, R indicates a range in which the true value is considered to exist around the observed value. P ^{(3 × 3)} _(-) is a prediction error covariance matrix of the track. Although omitted in Eqs. (10) and (11), the spread of the observation error covariance matrix and the spread of the track prediction error covariance matrix are also calculated in the present embodiment as in the first embodiment. is doing. As a result of the calculation of the equations (10) and (11), if the value on the right side of the equation (11) is smaller than 1, the bias error determination unit 34a determines that the observed value has a bias error and tracks the observed value. The transmission to the processing unit 20a is suspended. When the value on the right side of the equation (11) is 1 or more, the bias error determination unit 34a determines that the observed value has no bias error, and sends the observed value to the tracking processing unit 20a.

The node 100a performs the tracking process by removing the observed values having the bias error by performing the bias error determination process and generating a track only for the observed values necessary for maintaining the tracking accuracy. Therefore, the node 100a can suppress the deterioration of the tracking accuracy by performing the bias error determination process.

The configuration shown in the above-described embodiment shows an example of the content of the present invention, can be combined with another known technique, and is one of the configurations without departing from the gist of the present invention. It is also possible to omit or change the part.

1 target, 2 networks, 10,10a sensor, 20,20a tracking processing unit, 30 partial track unit, 31,31a prediction processing unit, 32 track fusion unit, 33 transmission adjustment unit, 34,34a bias error determination unit, 35th 1 calculation unit, 36 second calculation unit, 37 transmission unit, 40 integrated track unit, 41 integrated track generation unit, 42 integrated track storage unit, 50 communication device, 60, 60a display, 70 track storage unit, 100- 1-100-N, 100a node, 200 sensor network system, 300 control circuit, 300a processor, 300b memory.

Claims (6)

A sensor that observes the target and
A tracking processing unit that generates a partial wake of the target using the observed values of the sensor,
An integrated wake generator that generates an integrated wake using a plurality of the partial wakes,
A communication device that transmits the partial wake to the integrated wake generator,
A first spread of the covariance matrix of the prediction error of the integrated track using the difference between the predicted value of the integrated track and the smooth value of the partial track and the covariance matrix of the prediction error of the integrated track. The first calculation unit that calculates the calculation result of
Using the difference and the covariance matrix of the smoothing error of the partial track, a second calculation unit for calculating the second calculation result which is the spread of the covariance matrix of the smoothing error of the partial track, and
A bias error determination unit that determines whether to send the partial wake to the communication device by using the sum of the first calculation result and the second calculation result and the difference.
A target tracking device characterized by being equipped with. The target tracking device constitutes a sensor network system composed of a plurality of the target tracking devices.
The target tracking device according to claim 1, wherein the communication device transmits the partial track to another target tracking device constituting the sensor network system. The bias error determination unit
The quotient obtained by dividing the sum by the absolute value of the difference is calculated, and if the quotient is less than 1, the partial wake is suspended, and if the quotient is 1 or more, the partial wake is sent. The target tracking device according to claim 1 or 2. Information on the position of the partial wake is used for the smooth value.
The target tracking device according to any one of claims 1 to 3, wherein the predicted value uses information on the position of the integrated track. For the smooth value, information on the velocity of the partial wake is used.
The target tracking device according to any one of claims 1 to 3, wherein the predicted value uses information on the speed of the integrated wake. For the smooth value, information on the acceleration of the partial wake is used.
The target tracking device according to any one of claims 1 to 3, wherein the predicted value uses information on the acceleration of the integrated track.

Images