JP3373407B2 - Target 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 target tracking device for estimating target motion data such as a true value of a target position and a target velocity from detection data relating to target position information from a sensor or the like. .

[0002]

2. Description of the Related Art A conventional target tracking device will be described with reference to FIGS. 8 and 9 show, for example,
"Bar-Shalom Y. and Fortman TE:“ Tracking and
Data Association ”, Academic Press, New York (198
8) is a block diagram and a flowchart showing a configuration and an algorithm of the conventional target tracking device shown in FIG.

In FIG. 8, reference numeral 21 denotes a target observation device, and 22
Is a position error evaluator, 23 is a gain matrix calculator, 24 is a smoother, 25 is a predicted value calculator, 26 is a smoothing error evaluator, 2
7 is a prediction error evaluator, 28 is a delay circuit, and 29 is a delay circuit.

The target observing device 21 observes the target position information in polar coordinates, converts the result into target position information in rectangular coordinates, and outputs an observation vector.

In the position error evaluator 22, the predicted value calculator 2
The prediction vector obtained from step 5 before 1 sampling from the present time is input, the target position prediction vector is calculated, and the observation vector is obtained from the observation noise covariance matrix obtained from this and the preset observation model. A covariance matrix of the difference between the target position prediction vectors is calculated.

In the gain matrix calculator 23, an observation noise covariance matrix obtained from a preset observation model and 1
The gain matrix used for smooth vector calculation is calculated from the prediction error covariance matrix input through the delay circuit 29 from the prediction error evaluator 27 calculated before sampling.

The smoothing unit 24 determines the reliability of each observation vector obtained by the position error evaluator 22 and the observation vector obtained by the target observation apparatus 21, and determines the degree to which the observation vector contributes in smoothing vector calculation. The smoothing vector is calculated by applying the gain matrix obtained from the calculator 23.

The predictive value calculator 25 uses the smoothing vector obtained from the smoother 24 to predict constant velocity linear motion,
Sampling time t _{k + 1} one sampling after the present time
Compute the prediction vector for

In the smoothing error evaluator 26, the prediction error covariance matrix calculated one sampling before is inputted from the prediction error evaluator 27 through the delay circuit 29, and the gain matrix obtained by the gain matrix calculator 23 is used. , A smoothing error covariance matrix that is an evaluation value of the smoothing error of the smoothing vector is calculated.

The prediction error evaluator 27 uses the smoothing error covariance matrix obtained from the smoothing error evaluator 26 and the driving noise covariance matrix obtained from a preset motion model to sample the sampling time one sampling after the present time. t
_A prediction error covariance matrix, which is an evaluation of the prediction error of the prediction vector for _{k + 1} , is calculated.

Next, the operation of the above-described conventional target tracking device will be described with reference to FIG.

Based on the target position information obtained from the target observing device 21, the smoothed value of the target position and velocity and the initial value of the smoothed error covariance matrix are set based on the usual Kalman filter theory (ST1).

Next, in the predictive value calculator 25, the smoothed value from the smoother 24 is input, the uniform velocity linear motion prediction is performed, and the predicted value X _k (−) and the detection data in the next sampling are obtained. The matrix S _k for calculating the expected range (software gate) is expressed by equation (1),
It is calculated based on the equation (2) (ST2). At this time, Φ
_k-1 is a transition matrix, R _k is an observation noise covariance matrix, Γ ₂ (k) is an observation noise conversion matrix, and H is an observation matrix.

[0014]

[Equation 1]

[0015]

[Equation 2]

Next, in the prediction error evaluator 27, the prediction error covariance matrix for each motion model is calculated from the smoothing error covariance matrix input from the smoothing error evaluator 26 and the preset driving noise covariance matrix. Is calculated based on the equation (3) (ST3).

[0017]

[Equation 3]

Next, the detection data consisting of the position information which is the target signal detection result is input from the target observation device 21 (ST4). Among the obtained detection data, the one that satisfies the expression (4) is targeted for target tracking. Also, Z _k (−) is
The observable specifications are derived from the predicted value X _k (−), which is represented by the equation (5). At this time, assuming that the total number of obtained detection data is m _k , each detection data is expressed by Expression (6), and the entire detection data is expressed by Expression (7).

[0019]

[Equation 4]

[0020]

[Equation 5]

[0021]

[Equation 6]

[0022]

[Equation 7]

Next, in the position error evaluator 22, the normalized reliability β _k , _i that each detection data Z _k , _i is from the tracking target is calculated based on the equation (8) (ST).
5). At this time, the reliability γ _k , _i of each detection data is represented by Expression (9) and Expression (10). Further, P _D is the detection probability, P _GK is the probability that the detection data from the target enters the software gate, and β _k is the frequency per unit volume at which the unnecessary signal is detected at the sampling time t _k .

[0024]

[Equation 8]

[0025]

[Equation 9]

[0026]

[Equation 10]

Next, in the gain matrix calculator 23, the prediction error covariance matrix calculated one sampling before the current time is input from the prediction error evaluator 27 through the delay circuit 29, and the preset observation noise covariance matrix is set. The gain matrix K _k is calculated based on the equation (11) (ST6).

[0028]

[Equation 11]

Next, in the smoother 24, the reliability of each detection data is input, the detection data is input from the target observation device 21, the gain matrix is input from the gain matrix calculator 23, and the smoothed value is given by the equation ( 12) Calculate based on (ST)
7).

[0030]

[Equation 12]

Next, in the smoothing error evaluator 26, the prediction error covariance matrix is input from the prediction error evaluator 27, the gain matrix is input from the gain matrix calculator 23, and the smoothing error covariance matrix is expressed by the equation (13). Is calculated based on (ST8).
Here, ν _k , P ′ _k (+), ν _k and _i are respectively expressed by the formula (1
4), equation (15), and equation (16).

[0032]

[Equation 13]

[0033]

[Equation 14]

[0034]

[Equation 15]

[0035]

[Equation 16]

Then, this series of flow is repeated until the end of tracking (ST9).

[0037]

In the target tracking device based on the conventional Kalman filter theory as described above, at a certain time t, the position of the tracking target at the time t + Δt (Δt is a sampling interval) and the range in which the observation signal enters are determined. The target position is estimated by using all the observation signals that are predicted and fall within the range at time t + Δt.
In this method, when a plurality of signals are observed in a range in which the observation signal of the tracking target is predicted, the center of gravity obtained by weighting each signal with reliability is taken as the true target position. Therefore, if a new target occurs and the observation signal contains a signal from the new target, it will be processed as a signal from the already tracked target, and it will be delayed in handling the new target. There was a point.

Conventionally, there is a target tracking method using a multiple hypothesis. This method has a function of starting tracking for a new target, but the number of hypotheses increases in the course of tracking maintenance,
There is a problem that a processing load is applied.

The present invention has been made in order to solve the above-mentioned problems, and it is possible to increase the possibility of detecting the occurrence of a new target, and it is possible to accurately track the number of tracking targets together with the already tracked targets. The objective is to obtain a target tracking device that becomes

[0040]

[0041]

Means for Solving the Problems] The target tracking device according to this invention, the target tracking unit and multiple tentative based on Kalman filter
The switching between the predatory target tracking unit and the target unit based on predetermined conditions.
A tracking method switcher , wherein the tracking method switcher inputs the number of detection data from the target tracking unit based on the Kalman filter, and when the input number of detection data is more than a predetermined threshold, the target tracking process is performed by the Kalman filter. The target tracking unit based on the above is switched to the multiple hypothesis type target tracking unit.

Further, the target tracking device according to the present invention further comprises an estimated detection data number calculator for calculating the estimated detection data number based on the information obtained from the target tracking unit based on the Kalman filter, and the tracking method switching. The device inputs the number of detection data from the target tracking unit based on the Kalman filter, if the input number of detection data is large compared to the expected number of detection data, the target tracking process from the target tracking unit based on the Kalman filter. The target is switched to the multiple hypothesis type target tracking section.

Further, the target tracking device according to the present invention further comprises a counter for counting the number of samplings in which the number of detection data from the target tracking unit based on the Kalman filter is larger than a first predetermined threshold value, and the tracking method The switcher switches the target tracking processing from the target tracking unit based on the Kalman filter to the multiple hypothesis type target tracking unit when the cumulative value of the counter is larger than the second predetermined threshold value.

Further, the target tracking device according to the present invention further comprises an estimated detection data number calculator for calculating the estimated detection data number based on the information obtained from the target tracking unit based on the Kalman filter, and a target based on the Kalman filter. The number of detection data from the tracking unit is provided with a counter for measuring the number of times of sampling in which the number of detection data is larger than the expected number of detection data, and the tracking method switch is a target tracking process when the cumulative value of the counter is more than a predetermined threshold Is switched from the target tracking unit based on the Kalman filter to the multiple hypothesis type target tracking unit.

Further, the target tracking device according to the present invention further comprises a counter for measuring the number of samplings in which the number of detection data from the target tracking unit based on the Kalman filter is not obtained, and the tracking method switching device is the above-mentioned. When the cumulative value of the counter is larger than a predetermined threshold value, the target tracking process is switched from the target tracking unit based on the Kalman filter to the multiple hypothesis type target tracking unit.

Further, the target tracking device according to the present invention is
The tracking method switch, the false alarm probability is input from the target tracking unit based on the Kalman filter, if the input false alarm probability is greater than a predetermined threshold, the target tracking process from the target tracking unit based on the Kalman filter. This is to switch to the multi-hypothesis type target tracking unit.

[0047]

DETAILED DESCRIPTION OF THE INVENTION

Embodiment 1. Embodiment 1 of the present invention will be described with reference to FIG. 1 is a block diagram showing the configuration of a target tracking device according to Embodiment 1 of the present invention.
In each figure, the same reference numerals indicate the same or corresponding parts.

In FIG. 1, reference numeral 1 denotes a target tracking unit based on a Kalman filter which is equivalent to a target tracking method based on the conventional Kalman filter theory. Reference numeral 2 is a multiple-hypothesis type target tracking unit equivalent to the conventional target-tracking method based on the multiplex hypothesis disclosed in JP-A-8-271617. Further, 3 is a tracking method switch that switches between the target tracking unit 1 based on the Kalman filter and the multiple hypothesis type target tracking unit 2.

The tracking method switch 3 switches between the target tracking unit 1 based on the Kalman filter and the multiple hypothesis type target tracking unit 2 in accordance with a predetermined condition and the situation of the target number of tracking.

The target tracking device according to the first embodiment is
The tracking method switch 3 sets the initial values of the smoothed error covariance matrix in which the smoothed values such as the target position and the velocity and the error of the smoothed value are estimated, and then the predicted values such as the target position and the velocity and the error thereof are set. The estimated prediction error covariance matrix is calculated, the detection data is input from the target observation device 21 that outputs the signal detection result of the tracking target as the detection data, the gain matrix is calculated, and the smoothed values such as the target position and the velocity, and A smooth error covariance matrix that estimates the error is calculated, and a target tracking unit 1 based on a Kalman filter that repeats this series of processes until tracking is completed, and an area where detection data can exist from existing tracks included in each cluster is calculated. Calculate and select the tracking data corresponding to the track for each cluster from the possible existence area and all the detection data, and use the observation vector that does not correspond to the existing track without the cluster. Create a cluster, and if there is detection data that can be associated with tracks in multiple clusters, create a new cluster that integrates the corresponding clusters. Within each cluster, the detection data can be used as false signals, existing tracks, and new targets. The hypothesis for each case is generated, and the multiple hypothesis type target tracking unit 2 for extracting the target track from the hypothesis is switched and used.

As described above, according to the target tracking device of the first embodiment, it is possible to switch the target tracking method, and it is possible to perform accurate target tracking according to the situation of the tracking target number. .

Embodiment 2. A target tracking device according to Embodiment 2 of the present invention will be described with reference to FIG. FIG. 2 is a block diagram showing the configuration of the target tracking device according to the second embodiment of the present invention.

In FIG. 2, 1 and 2 are equivalent to those in the first embodiment. Further, 4 is a tracking method switcher based on the number of detection data, and when the number of detection data obtained at the sampling time is input from the target tracking unit 1 based on the Kalman filter at each sampling time, and it is larger than a threshold value given in advance. Determines that there is a high possibility that a new target has occurred, and switches the target tracking processing to the multiple hypothesis type target tracking unit 2.

At this time, the tracking method switching unit 4 sends information necessary for starting the target tracking based on the multiple hypothesis, such as the track information of the already tracked target, from the target tracking section 1 based on the Kalman filter to the multiple hypothesis type target tracking section 2. To deliver.

Further, the tracking method switching unit 4 switches the target tracking processing to the target tracking unit 1 based on the Kalman filter when the multi-hypothesis type target tracking unit 2 has established the track.

At this time, the tracking method switching unit 4 delivers the information necessary for the target tracking based on the Kalman filter such as the information of the established track from the multiple hypothesis type target tracking unit 2 to the target tracking unit 1 based on the Kalman filter. To do.

The target tracking device according to the second embodiment is
When the number of detection data obtained from the target observation device 21 of the target tracking unit 1 by the tracking method switching unit 4 exceeds a predetermined constant value, the Kalman filter-based target tracking unit 1 to the multiple hypothesis type target tracking unit Switch to 2.

According to the target tracking device of the second embodiment, the possibility of detecting the occurrence of a new target is increased by switching the target tracking method depending on the number of detection data, and tracking is performed together with the already tracked target. Accurate target tracking of the target number becomes possible.

Third Embodiment A target tracking device according to Embodiment 3 of the present invention will be described with reference to FIG. FIG. 3 is a block diagram showing the configuration of the target tracking device according to the third embodiment of the present invention.

In FIG. 3, 1 and 2 are the same as those in the first embodiment. Further, reference numeral 5 denotes a tracking method switcher based on the number of detection data, which inputs the number of detection data m _k obtained at the sampling time from the target tracking unit 1 based on the Kalman filter at each sampling time, and calculates the expected number of detection data. Expected number of detection data input from 6 m _{k p}
If there is a large number (m _k ≧ m _kp ), the target tracking processing is switched to the multiple-hypothesis target tracking section 2 as it is highly possible that a new target has occurred.

The predictive detection data number calculator 6 calculates the prediction detection data number m _kp according to the equation (17) based on the information obtained from the target tracking unit 1 based on the Kalman filter. Note that n _k is the number of already tracked targets, P _d is the detection probability, P _Gk is the probability that the target is inside the software gate, β ^k _FT is the number of unnecessary signals generated per unit volume, and V _k is the software. This is the volume of the wear gate.

_{M kp} = (n _k +1) · (P _d P _Gk ) + β ^k _FT V _k (17)

Information is passed between the target tracking unit 1 based on the Kalman filter and the multiple-hypothesis type target tracking unit 2 accompanying the switching of the tracking method, and from the multiple-hypothesis type target tracking unit 2 to the target tracking unit 1 based on the Kalman filter. The switching conditions are the same as those in the first embodiment.

The target tracking device according to the third embodiment is
The number of detection data obtained from the target observation device 21 of the target tracking unit 1 is predicted, and when the number of detection data exceeding this is obtained, the switching from the target tracking unit 1 based on the Kalman filter to the multiple hypothesis type target tracking unit 2 is performed. To do.

The target tracking device according to the third embodiment detects the occurrence of a new target according to the situation by calculating the optimum condition for switching the tracking method according to the target tracking environment for each sampling. Therefore, it is possible to accurately track the number of tracking targets together with the already-tracked targets.

Fourth Embodiment A target tracking device according to Embodiment 4 of the present invention will be described with reference to FIG. FIG. 4 is a block diagram showing the configuration of the target tracking device according to the fourth embodiment.

In FIG. 4, 1 and 2 are equivalent to those in the first embodiment. Further, 7 is a tracking method switcher based on the number of detection data, and when the number of detection data obtained at the sampling time is input from the target tracking unit 1 based on the Kalman filter at each sampling time, and the number is larger than a threshold value given in advance. Tells the counter 8 to increase the number of counts, inputs the cumulative number of counts from the counter 8, and if the number is larger than a preset threshold value, it is highly possible that a new target has occurred, and the multiple hypothesis The target tracking processing is switched to the type target tracking unit 2.

Further, the tracking method switching unit 7 instructs the counter 8 to reset the count number to 0 when the obtained detection data number is smaller than a predetermined threshold value.

That is, the counter 8 inputs an instruction from the tracking method switching unit 7 based on the number of detection data, and counts the number of samplings when the number of detection data is larger than a threshold value given in advance.

Information passing between the target tracking unit 1 based on the Kalman filter and the multiple-hypothesis type target tracking unit 2 accompanying the switching of the tracking method, and from the multiple-hypothesis type target tracking unit 2 to the target tracking device 1 based on the Kalman filter. The switching condition is
This is equivalent to the first embodiment.

The target tracking device according to the fourth embodiment is
The number of times that the number of detection data obtained from the target observation device 21 exceeds a predetermined fixed value is measured, and when the number of times occurs continuously for a predetermined number of times, the target tracking unit 1 based on the Kalman filter multi-hypothesis. Switch to the type target tracking unit 2.

According to the target tracking device of the fourth embodiment, frequent switching of the target tracking method can be suppressed under the unstable condition in which the observation condition of the sensor greatly changes for each sampling, and the stable operation is achieved. The target tracking device can be configured.

Embodiment 5. A target tracking device according to Embodiment 5 of the present invention will be described with reference to FIG. 5 is a block diagram showing the configuration of a target tracking device according to Embodiment 5 of the present invention.

In FIG. 5, 1 and 2 are equivalent to those of the above-mentioned first embodiment, 6 is equivalent to that of the above-mentioned third embodiment, and 8 is equivalent to that of the above-mentioned fourth embodiment.

Further, in FIG. 5, reference numeral 9 is a tracking method switching unit based on the number of detection data, and the number of detection data obtained at the sampling time is input from the target tracking unit 1 based on the Kalman filter at each sampling time to predict the value. When the number is larger than the expected number of detection data input from the detection data number calculator 6, the counter 8 is instructed to increase the count number, and the cumulative count number is input from the counter 8. When the number is larger than the threshold value given in advance, it is determined that a new target is likely to have occurred, and the target tracking processing is switched to the multiple hypothesis type target tracking unit 2.

Information is passed between the target tracking unit 1 and the multiple hypothesis type target tracking unit 2 based on the Kalman filter accompanying the switching of the tracking method, and from the multiple hypothesis type target tracking unit 2 to the target tracking unit 1 based on the Kalman filter. The switching conditions are the same as those in the first embodiment.

The target tracking device according to the fifth embodiment is
Predict the number of detection data obtained from the target observation device 21,
We measured the number of times we got more detection data than this,
When the number of times occurs consecutively a predetermined number of times, the target tracking unit 1 based on the Kalman filter is switched to the multiple hypothesis type target tracking unit 2.

According to the target tracking device of the fifth embodiment, frequent switching of the target tracking method can be suppressed under the unstable condition in which the observation condition of the sensor greatly changes for each sampling, and the stable operation can be achieved. The target tracking device can be configured.

Sixth Embodiment A sixth embodiment of the present invention will be described with reference to FIG. 6 is a block diagram showing the configuration of a target tracking device according to Embodiment 6 of the present invention.

In FIG. 6, 1 and 2 are equivalent to those in the first embodiment. Further, 10 is a tracking method switcher based on the number of detection data, which inputs the number of detection data obtained at the sampling time from the target tracking unit 1 based on the Kalman filter at each sampling time, and when the detection data cannot be obtained. , The counter 11 is instructed to increase the count number, the cumulative count number of the counter 11 is input, and when the number is larger than the threshold value given in advance,
The target tracking process is switched to the multiple hypothesis type target tracking unit 2 on the assumption that a new target is likely to occur.

Further, the tracking method switching unit 10 instructs the counter 11 to return the count number to 0 when the detection data number is obtained.

That is, the counter 11 inputs an instruction from the tracking method switching unit 10 based on the number of detection data, and counts the number of samplings when the number of detection data is not obtained.

Information passing between the target tracking unit 1 based on the Kalman filter and the multiple-hypothesis type target tracking unit 2 accompanying the switching of the tracking method, and from the multiple-hypothesis type target tracking unit 2 to the target tracking unit 1 based on the Kalman filter. The switching conditions are the same as those in the first embodiment.

The target tracking device according to the sixth embodiment is
The number of times that the detection data is not obtained from the target observation device 21 is measured, and when the number of times occurs continuously for a predetermined number of times, switching from the target tracking unit 1 based on the Kalman filter to the multiple hypothesis type target tracking unit 2 is performed. I do.

According to the target tracking device of the sixth embodiment, new detection data can be obtained even if the detection data cannot be obtained for a long time and the target track cannot be constructed due to the observation state of the sensor. When it becomes, it will be possible to start target tracking.

Seventh Embodiment A target tracking device according to Embodiment 7 of the present invention will be described with reference to FIG. 7 is a block diagram showing a configuration of a target tracking device according to Embodiment 7 of the present invention.

In FIG. 7, 1 and 2 are the same as those in the first embodiment. Further, reference numeral 12 is a tracking method switcher based on the false alarm probability, and the false alarm probability at the sampling time is input from the sensor of the target tracking unit 1 based on the Kalman filter for each sampling time, and if it is larger than a threshold value given in advance, The target tracking process is switched to the multiple hypothesis type target tracking unit 2 on the assumption that a new target is likely to occur.

Information passing between the target tracking unit 1 based on the Kalman filter and the multiple-hypothesis type target tracking unit 2 accompanying the switching of the tracking method, and from the multiple-hypothesis type target tracking unit 2 to the target tracking unit 1 based on the Kalman filter. The switching conditions are the same as those in the first embodiment.

The target tracking device according to the seventh embodiment is
When the false alarm probability obtained from the target observation device 21 (sensor) exceeds a predetermined threshold, the target tracking unit 1 based on the Kalman filter is switched to the multiple hypothesis type target tracking unit 2.

According to the target tracking device of the seventh embodiment, the occurrence of a new target may be detected even when the sensor performance is insufficient due to the false alarm probability of the sensor. Is set to a higher value, and it becomes possible to accurately track the number of tracking targets together with the already tracked targets.

[0091]

[0092]

Target tracking apparatus according to this invention exhibits, as described above, a target tracking unit based on a Kalman filter
Switch between multiple hypothesis type target tracking section based on predetermined condition
A tracking method switch, the tracking method switch inputs the number of detection data from the target tracking unit based on the Kalman filter, and if the input number of detection data is more than a predetermined threshold value, the target tracking processing is performed. Since the target tracking unit based on the Kalman filter is switched to the multi-hypothesis type target tracking unit, it is possible to increase the possibility of detecting the occurrence of a new target, and it is possible to accurately track the number of tracking targets together with the already tracked targets. Play.

Further, as described above, the target tracking device according to the present invention further comprises an estimated detection data number calculator for calculating the estimated detection data number based on the information obtained from the target tracking unit based on the Kalman filter. , The tracking method switch inputs the detection data number from the target tracking unit based on the Kalman filter, and if the input detection data number is larger than the expected detection data number, the target tracking process is performed on the Kalman filter. Since it switches from the target tracking unit based on the multiple hypothesis type target tracking unit, it is possible to increase the possibility of detecting the occurrence of a new target depending on the situation, and it is possible to accurately track the number of tracking targets together with the already tracked targets. Produce an effect.

Further, as described above, the target tracking device according to the present invention further comprises a counter for counting the number of samplings in which the number of detection data from the target tracking unit based on the Kalman filter is larger than the first predetermined threshold value. The tracking method switching device switches the target tracking processing from the target tracking unit based on the Kalman filter to the multiple hypothesis type target tracking unit when the cumulative value of the counter is larger than a second predetermined threshold value. In an unstable situation in which the observation status of 1 changes greatly with each sampling, it is possible to suppress frequent switching of the target tracking method and to configure a stable target tracking device.

As described above, the target tracking device according to the present invention further includes an estimated detection data number calculator for calculating the estimated detection data number based on the information obtained from the target tracking unit based on the Kalman filter. When the number of detection data from the target tracking unit based on the Kalman filter is a counter for measuring the number of sampling times more than the expected detection data number, the tracking method switching device, when the cumulative value of the counter is greater than a predetermined threshold value The target tracking process is switched from the target tracking unit based on the Kalman filter to the multi-hypothesis type target tracking unit.Therefore, in an unstable situation where the observation situation of the sensor changes significantly at each sampling, a frequent target tracking method is used. The effect that the switching can be suppressed and a stable target tracking device can be configured .

Further, as described above, the target tracking device according to the present invention further comprises a counter for measuring the number of samplings in which the number of detection data from the target tracking unit based on the Kalman filter is not obtained, and the tracking method The switcher switches the target tracking processing from the target tracking unit based on the Kalman filter to the multi-hypothesis type target tracking unit when the cumulative value of the counter is larger than a predetermined threshold value, and therefore, it may be detected for a long time depending on the observation state of the sensor. Even if the data cannot be obtained and the target track cannot be constructed, it is possible to start the target tracking when the detection data is newly obtained.

Further, the target tracking device according to the present invention is
As described above, the tracking method switch inputs the false alarm probability from the target tracking unit based on the Kalman filter, and if the input false alarm probability is larger than a predetermined threshold, the target tracking process is based on the Kalman filter. Since the target tracking unit switches to the multi-hypothesis type target tracking unit, even if the sensor performance is insufficient due to the state of false alarm probability of the sensor, it is highly possible to detect the occurrence of a new target. Therefore, it is possible to accurately track the number of tracking targets together with the already-tracked targets.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a target tracking device according to a first embodiment of the present invention.

FIG. 2 is a block diagram showing a configuration of a target tracking device according to a second embodiment of the present invention.

FIG. 3 is a block diagram showing a configuration of a target tracking device according to a third embodiment of the present invention.

FIG. 4 is a block diagram showing a configuration of a target tracking device according to a fourth embodiment of the present invention.

FIG. 5 is a block diagram showing a configuration of a target tracking device according to a fifth embodiment of the present invention.

FIG. 6 is a block diagram showing the configuration of a target tracking device according to a sixth embodiment of the present invention.

FIG. 7 is a block diagram showing the configuration of a target tracking device according to Embodiment 7 of the present invention.

FIG. 8 is a block diagram showing a configuration of a conventional target tracking device based on a Kalman filter.

FIG. 9 is a flowchart showing an algorithm of a target tracking device based on a conventional Kalman filter.

[Explanation of symbols]

1 Kalman filter based target tracking unit, 2 multiple hypothesis type target tracking unit, 3, 4, 5, 7, 9, 10 Tracking method switching unit based on the number of detection data, 6 Prediction detection data number calculator, 8, 11 Counter, 12 Tracking method switcher based on false alarm probability.

Continuation of the front page (56) Reference JP-A-9-90027 (JP, A) JP-A-9-251074 (JP, A) JP-A 2-116777 (JP, A) JP-A-59-195172 (JP , A) JP-A-6-138222 (JP, A) (58) Fields investigated (Int.Cl. ⁷ , DB name) G01S ⁷ /00-7/42 G01S 13/00-13/95

Claims (6)

(57) [Claims] 1. A tracking method switcher for switching a Kalman filter-based target tracking section and a multiple hypothesis type target tracking section based on a predetermined condition , wherein the tracking method switcher is based on the Kalman filter.
Enter the number of detection data from the target tracking unit,
Target tracking processing when the number of knowledge data is larger than a predetermined threshold
From the target tracking unit based on the Kalman filter to
A target tracking device characterized by switching to a multiple hypothesis type target tracking unit . 2.A target tracking unit based on the Kalman filter
Switch between multiple hypothesis type target tracking section based on predetermined condition
And tracking method switch Information obtained from the target tracking unit based on the Kalman filter
Prediction detection data to calculate the number of prediction detection data based on
With a number calculator, The tracking method switch is based on the Kalman filter
Enter the number of detection data from the target tracking unit,
When the amount of knowledge data is large compared to the number of expected detection data
The target tracking processing based on the Kalman filter
Switch from the tracking unit to the multiple hypothesis type target tracking unit thing
Characterized byEyesTarget tracking device. 3.A target tracking unit based on the Kalman filter
Switch between multiple hypothesis type target tracking section based on predetermined condition
And tracking method switch The detection data from the target tracking unit based on the Kalman filter
Number of samplings in which the number of data is greater than the first predetermined threshold
Equipped with a counter to measure In the tracking method switch, the accumulated value of the counter is the second
If more than the predetermined threshold value of
From the target tracking unit based on the Mann filter, the multiple hypothesis type
Switch to tracking unit Characterized byEyesTarget tracking device. 4.A target tracking unit based on the Kalman filter
Switch between multiple hypothesis type target tracking section based on predetermined condition
And tracking method switch Information obtained from the target tracking unit based on the Kalman filter
Prediction detection data to calculate the number of prediction detection data based on
A number calculator, The detection data from the target tracking unit based on the Kalman filter
The number of data samples is larger than the expected
With a counter that counts the number of times, The tracking method switch has a predetermined cumulative value of the counter.
If more than the threshold value of
The target tracking unit based on the filter from the multiple hypothesis type target tracking
Switch to department Characterized byEyesTarget tracking device. 5.A target tracking unit based on the Kalman filter
Switch between multiple hypothesis type target tracking section based on predetermined condition
And tracking method switch The detection data from the target tracking unit based on the Kalman filter
A cow that measures the number of samplings
And the The tracking method switch has a predetermined cumulative value of the counter.
If more than the threshold value of
The target tracking unit based on the filter from the multiple hypothesis type target tracking
Switch to department Characterized byEyesTarget tracking device. 6. A target tracking unit based on a Kalman filter
Switch between multiple hypothesis type target tracking section based on predetermined condition
And a tracking method switching device, wherein the tracking method switching device is based on the Kalman filter.
Enter the false alarm probability from the target tracking unit, and enter the false alarm
If the probability of alarm is greater than a predetermined threshold, target tracking processing
From the target tracking unit based on the Kalman filter, the multiplex
Goal tracker you and switches to the hypothesis type target tracking unit.