JP5183543B2 - Radar equipment - Google Patents

Description

  The present invention relates to a radar apparatus provided with tracking start means for detecting an unknown target in a radar apparatus applied to a multi-target tracking apparatus, and more particularly to a target observation value (hereinafter simply referred to as “observation value”). Included in the observation value so that the tracking start performance of the tracking start means is optimal when the observation data is regularly provided to the tracking start means by tracking the series data and generating the target track. The present invention relates to a radar apparatus that sets an observation interval that is one of the parameters to be measured.

In general, a radar apparatus requires a tracking technique for generating a target track using an observation value of position information obtained from a radar (or sensor) that observes a target.
For example, technologies for observing a target with a sensor and tracking the target using the obtained observation values have already been covered in many papers and patent documents, and various proposals have been made for tracking devices and tracking methods. Has been made.

  The tracking process is roughly divided into two types: a tracking start process and a tracking maintenance process. The tracking start process is a process for finding an undiscovered target from time-series data of observation values using a tracking function. On the other hand, the tracking maintenance process is a process of tracking the motion of a target whose existence is known.

  At this time, when tracking a target using a radar, the tracking maintenance process is known tracking information, so the direction in which the beam is directed is limited to some extent, while the tracking start process is Since the target is unknown, it is necessary to search with a beam directed to a wide area.

  Therefore, in the search of a wide area in the tracking start processing, in order to increase the target detection performance, it is better that the signal power is higher and the observation frequency is higher, so the search interval (observation interval) is short. The tracking start performance is higher.

  However, the transmission power per unit time of the radar is limited, and if the signal power is set to a high value, the observation interval must be lengthened. Conversely, if the observation interval is set to be short, the signal power is It must be lowered.

  Thus, there is a one-to-one relationship between the observation interval and the signal power. However, since the tracking start performance changes when the observation interval is changed, in order to effectively use the limited transmission power, It is necessary to control the radar so that the optimum observation interval is obtained.

The optimum observation interval of this type of radar can be determined to some extent at the time of radar design by performing a simulation in advance. However, since the optimal observation interval depends on the effective reflection area of the radar, the preset value is not optimal for an unexpected target.
Therefore, when an unexpected target becomes a target of tracking start, it is necessary to control the radar while estimating a new optimum observation interval.

Therefore, a resource management device that determines sensor resource allocation according to the tracking situation has been proposed as a radar device control technique using a sensor (see, for example, Patent Document 1).
In the conventional apparatus described in Patent Document 1, individual tracking situations of the tracking target track, such as the size of the error ellipse, the number of detection omissions, the target distance, the target speed, and the movement parameters such as the distance between multiple targets, etc. Accordingly, priorities are set, and sensor resources are preferentially assigned to a target track having a high priority.

FIG. 7 is a block diagram showing a conventional device described in Patent Document 1, and shows a functional configuration when applied to sensor control for realizing tracking maintenance optimization.
Hereinafter, the operation of the conventional apparatus shown in FIG. 7 will be described by limiting only to the sensor control function.

In addition, although the above-mentioned patent document 1 also describes the control of parameters that determine the operation of the tracking filter, the control function of the tracking filter is not used here.
Further, in Patent Document 1, it is assumed that observation conditions such as information on a clutter occurrence area can be obtained. However, information on a clutter occurrence area cannot be obtained here.
That is, here, only the function for determining sensor assignment based on the tracking situation will be described. Therefore, in FIG. 7, there are blocks that are not used in the following description.

  7, the resource management device of the radar device includes a sensor group 100 including a plurality of sensors 100a, 100b,..., A multi-target tracking processing unit 101 having a tracking start function, a sensor group instruction unit 102, and a sensor. An observation information fusion unit 103 that generates observation values and observation information based on information from the group 100, a filter group instruction unit 104, a sensor state extraction unit 105, a target state evaluation unit 106, an observation information extraction unit 107, A tracking information extraction unit 108, a filter state extraction unit 109, a resource allocation method creation unit 110 having an additional database, and a resource management calculation unit 111 having a resource database are provided.

The multi-target tracking processing unit 101 includes a correlation determining unit 101a having a combined wake correlation matrix and a filter processing unit 101b having a tracking filter (not shown).
The correlation determination unit 101 a generates a combination of the observation value and the wake based on the observation value from the observation information fusion unit 103.

  The filter processing unit 101b controls the tracking information extraction unit 108 and the filter state extraction unit 109 using the observation value from the observation information fusion unit 103 and the combination of the observation value and the wake from the correlation determination unit 101a as input information. .

The sensor group 100 observes the designated target at the designated time according to the operation condition output by the sensor group instruction unit 102, and outputs the observation information.
When observation information is input from any of the sensors in the sensor group 100, the observation information fusion unit 103 receives an observation value (target observation) for a tracking filter that estimates the motion specifications of the target based on the operation condition. Value).

The tracking filter performs tracking calculation of the corresponding target, thereby estimating and predicting the position and speed of each target and updating the tracking information of the target.
The tracking information extraction unit 108 uses an error ellipse (error covariance matrix) that is an estimation error range as information related to an estimated value of a target motion specification (for example, position and velocity) from target tracking information output by the tracking filter group. ) And the extracted information is input to the target state evaluation unit 106.

  The target state evaluation unit 106 sets an evaluation value corresponding to the priority of sensor allocation for each target. Since an estimation error range is obtained from the tracking filter, a large evaluation value is set for the purpose of reducing the error for a target having a wide estimation error range.

  Next, when the evaluation value from the target state evaluation unit 106 is input, the resource allocation method creation unit 110 refers to the tracking database and, based on the evaluation value output from the target state evaluation unit 106, for each target. A plurality of sets of sensor assignments and evaluation values for the distribution effects are calculated and output.

That is, based on the evaluation value (the tracking state of each target) output by the target state evaluation unit 106, the sensor and observation time assigned to the target are determined.
This allocation is generated as a plurality of candidates, and for each, the distribution effect evaluation value obtained by weighting the expected value of how much the tracking accuracy of each target is improved by observation of the allocated sensor by the size of the target state evaluation value Is calculated.

  Next, the resource management calculation unit 111 receives a plurality of sensor assignments and distribution effect evaluation values for each resource from the resource allocation method creation unit 110, and inputs the current operation status of the sensor group 100. While referring to the database, considering whether each sensor allocation is feasible and to what extent the load such as radiant energy applied to the sensor group 100 reaches, sensor allocation with a large distribution effect evaluation value is selected.

  When the resource management calculation unit 111 determines an optimal distribution method, the sensor group instruction unit 102 issues an operation instruction to each sensor 100a, 100b,... Constituting the sensor group 100 according to the determination information.

  The sensor control procedure of the resource management apparatus in the conventional apparatus is as described above. In the conventional apparatus of FIG. 7, at the time when the observation value by the sensor group 100 is input to the tracking means, which observation value becomes which tracking target. It is premised on that the response is known to some extent, and the allocation of sensors is determined based on the estimation error of each tracking target.

That is, sensor resources are preferentially allocated for targets with relatively large errors and for combinations of targets and sensors that have a relatively high improvement effect by sensor observation.
For example, as in the example shown in the explanatory diagrams of FIGS. 8A and 8B, when the error ellipses of the targets O1 and O2 are compared and the target O2 is larger than the target O1, Sensor resources are preferentially allocated to the target O2.

Japanese Patent No. 4014785

When a conventional radar device starts tracking while searching for a specific area by a radar, the tracking start is a short-time process in which the first few observation values are obtained, and the estimation error of the wake is not converged. It is insufficient.
Therefore, when trying to execute the control to optimize the observation interval, the method of controlling the observation interval while actually tracking is not in time, so the observation interval is optimized before actually performing the observation with the radar. However, there is a problem that the request cannot be realized.

  The present invention has been made to solve the above-described problems, and is for optimizing the tracking start performance so as to efficiently improve the estimation accuracy of the wake generated by the tracking technique. An object of the present invention is to obtain a radar apparatus capable of determining the optimum observation interval of the radar.

  A radar device according to the present invention generates a target track by tracking a time series data of observation values obtained by a radar that searches a predetermined observation area, a radar control unit that controls the radar, and an unknown value. Tracking start means for detecting a target, and a radar apparatus that periodically provides an observation value from the radar to the tracking start means. An observation interval that is one of the parameters of the observation value is set as an actual value in the tracking start means. Simulation processing means for setting the tracking start performance to an optimum value is further provided. The simulation processing means searches for an observation interval in performance analysis and sets observation interval candidates, and generates an optimum observation interval predicted value. Estimated tracking start performance value based on the observation interval candidate set by the observation interval setting unit for analysis and the observation interval setting unit for analysis Tracking start performance calculator that analytically predicts based on specific assumptions related to wake generation from radar observation performance values and simulation observation interval candidate setting that searches observation intervals in simulations and sets candidate observation intervals And a pseudo-observation value generator for generating a pseudo-observation value used for the tracking start simulation based on the observation performance value of the radar, and the pseudo-observation value generated by the pseudo-observation value generator as input information. A pseudo wake generation unit that generates a pseudo wake with a tracking processing function equivalent to the tracking start means that provides the tracking start performance, and a tracking start performance totaling unit that calculates and calculates the tracking start performance using the pseudo wake generated by the pseudo wake generation unit; And a radar parameter instruction unit for instructing the radar control unit to set the observation interval determined by the simulation. It is intended.

  According to the present invention, it is possible to determine the optimal observation interval of the radar for optimizing the tracking start performance so that the improvement in estimation accuracy of the wake generated by the tracking technique is efficiently realized.

It is a block diagram which shows the structure of the radar apparatus which concerns on Embodiment 1 of this invention. It is a flowchart which shows the process sequence for the observation interval determination of the radar by Embodiment 1 of this invention. It is explanatory drawing which shows the example of the tracking start performance estimated value according to the observation interval candidate in Embodiment 1 of this invention. It is explanatory drawing which shows the example of the 1st tracking start performance according to the observation interval candidate in Embodiment 1 of this invention. It is explanatory drawing which shows the example of the 2nd tracking start performance according to the observation interval candidate in Embodiment 1 of this invention. It is explanatory drawing which shows the example of the 3rd tracking start performance according to the observation interval candidate in Embodiment 1 of this invention. It is a block diagram which shows the principal part structure of the conventional radar apparatus. It is explanatory drawing which shows each error ellipse of a different target in the conventional radar apparatus.

Embodiment 1 FIG.
Hereinafter, preferred embodiments of the present invention will be described with reference to the drawings.
FIG. 1 is a block diagram showing a radar apparatus according to Embodiment 1 of the present invention, and shows a schematic configuration when applied to a multi-target tracking apparatus.

  In FIG. 1, the radar apparatus includes an observation condition input unit 1 serving as a user input interface, an analysis observation interval setting unit 2, a tracking start performance calculation unit 3, a simulation observation interval candidate setting unit 4, and pseudo observation. A value generation unit 5, a pseudo wake generation unit 6, a tracking start performance totaling unit 7, a radar parameter instruction unit 8, a radar control unit 9, and a radar 10 are provided.

The observation condition input unit 1 generates a relationship between the observation interval and the signal power according to the target type input from the user and inputs the relationship to the observation interval setting unit 2 for analysis.
Based on the input information from the observation condition input unit 1 and the feedback information from the tracking start performance calculation unit 3, the analysis observation interval setting unit 2 generates an observation interval candidate and an optimum observation interval expected value Tpre, The interval candidate is input to the tracking start performance calculation unit 3, and the optimum observation interval predicted value Tpre is input to the simulation observation interval candidate setting unit 4.

The tracking start performance calculation unit 3 generates a predicted tracking start performance value based on the observation interval candidate, and feeds it back to the analysis observation interval setting unit 2.
The simulation observation interval candidate setting unit 4 generates an observation interval and an optimal observation interval based on the input information from the analysis observation interval setting unit 2 and the feedback information from the tracking start performance totaling unit 7, and The interval is input to the pseudo observation value generation unit 5, and the optimum observation interval is input to the radar parameter instruction unit 8.

The pseudo observation value generation unit 5 inputs a pseudo observation value based on the observation interval to the pseudo track generation unit 6, and the pseudo track generation unit 6 inputs the pseudo track based on the pseudo observation value to the tracking start performance totaling unit 7.
The tracking start performance totaling unit 7 feeds back the tracking start performance based on the pseudo track to the simulation observation interval candidate setting unit 4.

The radar parameter instruction unit 8 inputs an observation interval based on the optimum observation interval to the radar control unit 9.
Finally, the radar control unit 9 controls the radar 10 based on the observation interval from the radar parameter instruction unit 8.

  The radar 10 is a part included in a conventional radar apparatus and has the same configuration as described above (see FIG. 7). That is, instead of the sensor group 100 in FIG. 7, only the point having a radar is different from that in FIG. 7, and therefore, the specific configuration of the multi-target tracking processing unit 1 and the like is not shown here and is generic. In FIG. That is, the radar 10 in FIG. 1 has a multi-target tracking function and a tracking start function.

  Each block in FIG. 1 is a radar apparatus that periodically provides observation values from the radar 10 to the tracking start means, and sets the observation interval, which is one of the parameters of the observation values by the radar 10, in the tracking start means. The simulation processing means for setting the tracking start performance to the optimum value is configured.

  In the radar apparatus shown in FIG. 1, the optimum observation interval of the radar 10 is determined on the premise that the tracking of a target is started while searching for a certain area by one radar 10. Is instructed to the radar control unit 9.

The processing contents according to the first embodiment of the present invention shown in FIG. 1 will be specifically described below with reference to FIGS.
FIG. 2 is a flowchart showing a processing procedure for determining the observation interval of the radar 10 by the radar apparatus of FIG.

FIG. 3 is an explanatory diagram showing predicted tracking start performance values corresponding to three observation interval candidates, and FIGS. 4 to 6 are explanatory diagrams showing first to third tracking start performance values corresponding to three observation intervals. FIG.
Here, it is assumed that a certain target is a tracking start target, and the optimal observation interval is not estimated in advance for that type of target.

  In FIG. 2, first, observation condition input (step S1) is performed by a user operation, and the observation condition input unit 1 receives the target effective reflection area σ to be tracked and the transmission power Ptr of the radar 10 as follows. Input as observation conditions.

At this time, the observation condition input unit 1 obtains a relational expression between the signal power SNR and the observation interval T based on the input information of the observation condition.
The relationship between the signal power SNR and the observation interval T is expressed by the following equation (1) using the transmission power Ptr of the radar 10 and the target effective reflection area σ.

  SNR · T = f (Ptr, σ) (1)

However, in equation (1), the function f () is uniquely determined by the radar 10.
Next, the analysis observation interval setting unit 2 increases the observation interval candidate by a predetermined step size ΔT (a fixed parameter set in advance) from the previously set observation interval candidate based on the equation (1). And input to the tracking start performance calculation unit 3 (step S2).

  When step S2 is first executed, the observation interval candidate is set to a minimum value Tmin that can be taken, and the minimum observation interval candidate is input to the tracking start performance calculation unit 3. The minimum value Tmin is a parameter set in advance. Hereinafter, the observation interval candidate is increased by the step size ΔT each time Step S2 is repeatedly executed.

Next, the tracking start performance calculation unit 3 calculates a predicted tracking start performance value when the observation interval candidate set in step 2 is set as an observation interval (step S3).
At this time, the start probability Ptgt of the target track is applied as the predicted tracking start performance value calculated by the tracking start performance calculation unit 3.

Hereinafter, a procedure for deriving the start probability Ptgt of the target wake will be specifically described.
First, at the end of the tracking process at the observation time t (i), the probability pm (i) that the target track is in the m (m = 0, 1, 2, 3) detection state is calculated.

This probability pm (i) represents the probability that m observed values of the target wake are obtained and the m observed values constitute one target wake.
The probability pm (i) (m = 0, 1, 2, 3) is described by the following equation (2) consisting of a recurrence formula using the probability pm (i−1) one observation time before.

p0 (i) = p0 (i-1). {1-pD (i)} + p2 (i-1). {1-pD (i)}. pF (i)
p1 (i) = p1 (i-1). {1-pD (i)} + p0 (i-1) .pD (i)
p2 (i) = p2 (i-1). {1-pD (i)}. {1-pF (i)} + p1 (i-1) .pD (i)
p3 (i) = p3 (i-1) + p2 (i-1) .pD (i)
... (2)

In equation (2), pD (i) is the probability that the radar will detect the target at the observation time t (i), and is a value determined by the signal power SNR.
Further, pF (i) is the probability that a false alarm will enter the gate of the target track. If the power is represented by “^”, it is calculated as the following equation (3).

  pF (i) = 1- (1-pfa) ^ [Vg (i) / {ΔR · R (i) · ΔE1 · R (i) · ΔAz}] (3)

However, in formula (2), pfa is a false alarm probability determined by the setting of the radar detection threshold.
Vg (i) is the gate volume of the target track and is calculated by tracking filter calculation. R (i) is a target distance, and ΔR, ΔE1, and ΔAz are resolutions in the radar distance direction, elevation angle direction, and azimuth angle direction, respectively.

  From the equation (2), the initial value of the probability pm (i) that the target track is in the m detection state, that is, the value at “i = 1” is the following equation (4).

p0 (1) = 1-pD (1)
p1 (1) = pD (1)
p2 (1) = 0
p3 (1) = 0
... (4)

  Here, assuming that the target track has three detection states, that is, when the three observation values are obtained, the target track start probability Ptgt (i) at the observation time t (i) is It is expressed as equation (5).

  Ptgt (i) = p3 (i) (5)

Returning to FIG. 2, following step S3, the tracking start performance calculation unit 3 determines whether a maximum value of the tracking start performance has been found (step S4).
That is, it is determined whether or not the maximum value of the tracking start performance predicted value (target track start probability Ptgt) calculated as in Expression (5) is found.

In step S4, when the maximum value is not found (the maximum value is not fixed), the determination result is “No”, and the process returns to the analysis observation interval setting process (step S2).
In step S2, the observation interval is increased by increasing the step size ΔT from the current observation interval candidate value.

  Thereafter, the above processing (steps S2 to S4) is repeated, and in step S4, when the maximum value of the predicted tracking start performance value is determined and the determination result is “Yes”, the next observation interval candidate setting process for simulation is performed. The process proceeds to (Step S5).

  Whether or not the maximum value of the predicted tracking start performance value has been determined is determined based on whether the current (latest) predicted tracking start performance value is the tracking start performance obtained by the tracking start performance calculation performed one time before. The determination can be made based on whether or not the value is lower than the value.

Here, the determination process in step S4 will be specifically described.
In the example of the tracking start performance predicted value according to the observation interval candidate shown in FIG. 3, the tracking start performance predicted value when the observation interval candidate is the first “Tmin” and the second “Tmin + ΔT”. In comparison with the predicted tracking start performance value, the second tracking start performance predicted value is larger.

Therefore, the maximum value of the expected tracking start performance value is not fixed at the second time point.
Next, when the observation interval candidate is set to “Tmin + 2ΔT” for the third time, the tracking start performance predicted value is lower than the second value, and therefore the tracking start performance prediction is performed in the third search. The observation interval candidate that realizes the maximum value is determined to be “Tmin + ΔT”.
This maximum value is set as the maximum value of the predicted tracking start performance value obtained so far.

  Thus, when the maximum value is determined, an observation interval candidate that maximizes the tracking start performance predicted value is input to the simulation observation interval candidate setting unit 4 as the optimum observation interval predicted value Tpre, and the next simulation observation interval is set. The process proceeds to candidate setting processing (step S5).

In step S5, observation interval candidates for simulation are set.
It should be noted that when step S5 is executed for the first time after the optimum observation interval expected value Tpre is input from the analysis observation interval setting unit 2, three observation intervals “Tpre−ΔT” and “Tpre” , “Tpre + ΔT” is an observation interval candidate for simulation.

Next, the pseudo-observation value generation unit 5 generates N pseudo-observation values while simulating actual observation conditions with random numbers using the observation interval candidates set in step S5 as observation intervals (step S6). .
The number N of pseudo-observation values generated is the number of trials of Monte Carlo simulation (a parameter set in advance), which is the number of times that can be spent for simulation.

Next, the pseudo wake generation unit 6 uses the pseudo observation value generated by the pseudo observation value generation unit 5 as input information, and performs the pseudo processing by the tracking processing function equivalent to that of the tracking start means in which the radar 10 provides the observation value. A wake is generated (step S7).
This pseudo wake generation process (step S7) is executed for all of the N pseudo observation values.

  Next, the tracking start performance totaling unit 7 performs a simulation result totaling process, selects only the target track from the pseudo track generated by the pseudo track generating unit 6, and sets the target at a specific time t (i). The wake start probability PT (i) is calculated (step S8).

  At this time, the tracking start performance totalization unit 7 determines whether the wake generated by the pseudo wake generation unit 6 is a target wake or a false wake. That is, if the wake generated by the quasi-wake generation unit 6 is only the quasi-observation value previously used for motion specification estimation assuming the target, it is determined as the target wake.

On the other hand, if the wake generated by the quasi-wake generation unit 6 includes a false observation value assumed in the past in the pseudo-observation value, the tracking start performance totaling unit 7 indicates that the generated wake is an erroneous wake. judge.
Here, if the number of simulation trials in the pseudo-observation value generation process (step S6) is N, and the target track is established by the mT trial by the observation time t (i), the target track start probability PT (i ) Is expressed as the following equation (6).

  PT (i) = mT / N (6)

  Next, the tracking start performance totaling unit 7 determines whether or not the maximum value of the tracking start performance (start probability PT (i)) obtained in step S8 has been found (step S9), and the maximum of the tracking start performance is determined. When the value is obtained and the determination result is “Yes”, the process proceeds to the last radar control instruction process (step S10).

  On the other hand, if the maximum value of the tracking start performance is not obtained in step S9, the determination result is “No”, the process returns to step S5, the simulation observation interval candidate setting process is performed again, and the above process (step S5 to S9) are repeatedly executed.

Here, the simulation interval candidate setting process in steps S5 to S9 will be specifically described.
As described above, when the step S5 is first executed, the three observation intervals “Tpre−ΔT”, “Tpre”, and “Tpre + ΔT” are the observation interval candidates for simulation. Therefore, the simulation result totaling process (step S8) In FIG. 4, for example, three tracking start performance values are obtained as shown in FIG.

  In FIG. 4, the tracking start performance with respect to three observation intervals has an upwardly convex characteristic. Thus, when three tracking start performances are obtained, the tracking start performance at the center candidate “Tpre” is the highest, and the maximum value of the tracking start performance is realized at this observation interval “Tpre”. The observation interval “Tpre” is determined as the optimum observation interval.

On the other hand, as shown in FIG. 5, when the tracking start performance for the three observation intervals shows a monotonic increase, the tracking start performance at the third observation interval “Tpre + ΔT” is the highest, but this observation interval “Tpre + ΔT” It is not possible to confirm that the maximum has been realized.
Therefore, the observation interval candidate is further increased by the increment ΔT, and the observation interval that achieves the maximum of the tracking start performance is searched while sequentially setting “Tpre + 2 · ΔT”, “Tpre + 3 · ΔT”,. To do.

Further, as shown in FIG. 6, when the tracking start performance for the three observation intervals shows a monotonic decrease, the performance at the first observation interval “Tpre−ΔT” is the highest, but this observation interval “Tpre− It cannot be determined that the maximum is realized by “ΔT”.
Therefore, the maximum of the tracking start performance is realized while further reducing the observation interval candidates by increments of ΔT and sequentially setting “Tpre-2 · ΔT”, “Tpre-3 · ΔT”,. Search for observation intervals.

Returning to FIG. 2, finally, the radar parameter instruction unit 8 instructs the radar control unit 9 to perform radar control (step S <b> 10), and the optimal observation obtained while searching for the observation interval of the radar 10 by simulation. Instructions are output so that there is an interval.
In this way, the observation interval can be optimally determined by analysis and simulation before actually operating the radar 10 and performing observation.

  As described above, the radar apparatus according to Embodiment 1 of the present invention includes the radar 10 that searches for a predetermined observation region, the radar control unit 9 that controls the radar 10, and the time-series data of the observation values obtained by the radar 10. A tracking start unit for detecting an unknown target by generating a target track by tracking processing; and a radar apparatus that periodically provides an observation value from the radar to the tracking start unit. The apparatus further includes simulation processing means for setting one observation interval to a value at which the actual tracking start performance in the tracking start means is optimum.

  The simulation processing means includes an analysis observation interval setting unit 2, a tracking start performance calculation unit 3, a simulation observation interval candidate setting unit 4, a pseudo observation value generation unit 5, a pseudo track generation unit 6, and a tracking start performance. A totaling unit 7 and a radar parameter instruction unit 8 are provided.

The analysis observation interval setting unit 2 searches for observation intervals in performance analysis, sets observation interval candidates, and generates an optimum observation interval expected value Tpre.
The tracking start performance calculation unit 3 calculates the predicted tracking start performance value based on the observation interval candidate set by the analysis observation interval setting unit 2 based on a specific assumption regarding the wake generation from the observation performance value of the radar 10. Analytical prediction calculation.

The simulation observation interval candidate setting unit searches for an observation interval in the simulation and sets a candidate observation interval.
The pseudo observation value generation unit 5 generates a pseudo observation value used for the tracking start simulation based on the observation performance value of the radar 10.

The pseudo wake generation unit 6 uses the pseudo observation value generated by the pseudo observation value generation unit 5 as input information, and generates a pseudo wake by a tracking processing function equivalent to the tracking start means in which the radar 10 provides the observation value.
The tracking start performance totaling unit 7 calculates and calculates the tracking start performance using the pseudo track generated by the pseudo track generating unit 6, and the radar parameter instruction unit 8 performs radar control on the set value of the observation interval determined by the simulation. The unit 9 is instructed.

Thereby, the optimal observation interval of the radar 10 for optimizing the tracking start performance can be efficiently determined so that the estimation accuracy of the wake generated by the tracking technique is improved.
Therefore, even when tracking is started for a target that has not been tracked, the start of tracking can be optimized by setting an observation interval determined by a short analysis and simulation.

The predicted tracking start performance value calculated by the tracking start performance calculation unit 3 and the tracking start performance calculated by the tracking start performance totaling unit 7 are the start probability of the target track at a specific time t (i). PT (i).
Thus, by applying the target track start probability PT (i) as the tracking start performance index, it is possible to set the observation interval for the highest target track start probability PT (i). Become.

  In addition, since the number of searches in the simulation is minimized with the optimum observation interval predicted by the analysis as a starting point, not the search by simply repeating the simulation alone, the observation interval can be determined in a realistic time.

Embodiment 2. FIG.
In the first embodiment, the expected tracking start performance value calculated by the tracking start performance calculation unit 3 and the tracking start performance calculated by the tracking start performance totaling unit 7 are set to a specific time t (i ), The start probability PT (i) of the target wake may be used instead of the start probability PT (i) of the target wake.

Hereinafter, a radar apparatus according to Embodiment 2 of the present invention, in which the erroneous track start probability PF is applied to the multi-target tracking apparatus, will be described.
The apparatus configuration according to the second embodiment of the present invention is the same as that described above (see FIG. 1), and the processing procedure is also as shown in FIG.

  In this case, however, the analysis observation interval candidate setting process (step S2), the tracking start performance calculation process (step S3), the simulation observation interval candidate setting process (step S5), and the simulation result totaling process (step S8) in FIG. ) Is partially different, so that the following description will be limited to steps S2, S3, S5, and S8.

First, in step S <b> 2, the analysis observation interval setting unit 2 searches for an observation interval so that the “number of false wakes” started in the tracking start process is as small as possible.
Subsequently, in step S3, the tracking start performance calculation unit 3 calculates a tracking start performance expected value when the observation interval candidate input in step S2 is set as an observation interval.

At this time, the tracking start performance calculation unit 3 applies the erroneous track start probability Pfls as the calculated tracking start performance predicted value.
Here, when the expected value of the number of false wakes in the m (m = 1, 2, 3) detection state at the end of the tracking process at the observation time t (i) is represented by nFm (i), each expected value nFm (i ) Is described as the following equation (7) consisting of a recurrence formula using the expected value nFm (i−1) one observation time ago.

nF1 (i) = nF1 (i-1) · {1-pfalse (i)} + pfa · Vobs / {ΔR · R (i) · ΔE1 · R (i) · ΔAz}
nF2 (i) = nF2 (i-1). {1-pfalse (i)} + nF1 (i-1) .pfalse (i)
nF3 (i) = nF3 (i-1) + nF2 (i-1) .pfalse (i) + p2 (i-1). {1-pD (i)}. pF (i)
... (7)

  However, in Formula (7), Pfalse (i) is the probability that a false alarm will be given to any of the resolution cells in the gate {volume ΔR · R (i) · ΔE1 · R (i) · ΔAz}, This can be calculated by subtracting the probability that no false alarm will occur in all cells from “1.0”. Accordingly, the probability Pfalse (i) that a false alarm is given to any one of the resolution cells in the gate is calculated as the following equation (8) when the power is represented by “^”.

  Pfalse (i) = 1− (1-pfa) ^ [Vfg (i) / {ΔR · R (i) · ΔE1 · R (i) · ΔAz}] (8)

However, in Expression (8), the gate volume Vfg (i) is the upper limit value of the gate volume, and Vobs is the volume of the search area.
Further, the expected value (that is, initial value) nFm (i) of the number of erroneous wakes in each state at the target first detection time is expressed as the following equation (9).

nF1 (i) = pfa · Vobs / {ΔR · R (i) · ΔE1 · R (i) · ΔAz}
nF2 (i) = 0
nF3 (i) = 0
... (9)

  Here, assuming that an erroneous track is established by three observation values, the expected value nF3 (i) at the first detection time obtained from Equation (9) is divided by the cell number ncell of the resolution of the entire observation region. The value is the false track start probability Pfls (i), and is expressed as the following equation (10).

  Pfls (i) = nF3 (i) / ncell (10)

  Hereinafter, the simulation observation interval candidate setting unit 4 according to the second embodiment of the present invention searches for an observation interval for minimizing the value of the erroneous track start probability Pfls (i) in step S5. That is, the simulation observation interval candidate setting unit 4 searches for the observation interval so that the number of false wakes started in the tracking start process is as small as possible.

  Further, the tracking start performance totaling unit 7 selects only a false track from the pseudo track generated by the pseudo track generating unit 6 in the simulation result totaling process (step S8), and selects a specific time t (i). The start probability PF of the false wake at is calculated.

  Here, assuming that the number of simulation trials is N and a total of mF mistracks have been established in all trials by the observation time t (i), the mistrack start probability PF (i) is expressed by the following equation: It is expressed as (11).

  PF = mF / (N · ncell) (11)

  As described above, according to the second embodiment of the present invention, the expected tracking start performance value calculated by the tracking start performance calculation unit 3 and the tracking start performance calculated by the tracking start performance totaling unit 7 are: Since the erroneous track start probability PF (i) is applied as the tracking start performance index PF (i) at the specific time t (i), the erroneous track start probability PF is the lowest. It is possible to efficiently set the observation interval.

Embodiment 3 FIG.
In the second embodiment, the tracking start performance predicted value calculated by the tracking start performance calculation unit 3 and the tracking start performance calculated by the tracking start performance totaling unit 7 are set at a specific time t (i However, instead of the erroneous track start probability PF (i), virtual signal-to-noise ratios VSNRp and VSNR required by the radar 10 may be used.

Hereinafter, a radar apparatus according to Embodiment 3 of the present invention in which a virtual signal-to-noise ratio is applied to a multi-target tracking apparatus will be described.
The apparatus configuration according to the third embodiment of the present invention is the same as that described above (see FIG. 1), and the processing procedure is also as shown in FIG.

  However, in this case, since the contents of the tracking start performance calculation process (step S3) and the simulation result totaling process (step S8) in FIG. 2 are partially different, only the steps S3 and S8 will be described below.

First, the tracking start performance calculation unit 3 calculates a predicted tracking start performance value in step S3 when the observation interval candidate input in step S2 is set as an observation interval.
At this time, the tracking start performance calculation unit 3 determines that the start probability Ptgt (i) and Pfls (i) of the target track and the erroneous track at a specific time t (i) are the detection probability and false alarm probability of the radar 10. And a virtual signal-to-noise ratio VSNRp that realizes the detection probability and the false alarm probability is set as a predicted tracking start performance value.
The virtual signal-to-noise ratio VSNRp required by the radar 10 is calculated as in the following equation (12).

VSNRp = 10 · log ₁₀ [log {Ptgt (i)} / log {Pfls (i)} − 1] (12)

  In the equation (12), the calculation method of the target track start probability Ptgt (i) and the erroneous track start probability Pfls (i) is as shown in the above-described equations (5) and (10), respectively.

  Further, the tracking start performance totaling unit 7 classifies the pseudo wake generated by the pseudo wake generating unit 6 into a target wake and a false wake in the simulation result totaling process (step S8), and sets each start probability of the radar 10 A virtual signal-to-noise ratio VSNR that realizes the detection probability and the false alarm probability is calculated assuming that it corresponds to the detection probability and the false alarm probability.

  Here, assuming that the number of simulation trials is N, the target track is established by mT trials by the observation time t (i), and a total of mF false wakes are established by all trials, the virtual signal pair The noise ratio VSNR is calculated as in the following equation (13).

VSNR = 10 · log ₁₀ {log (PT) / log (PF) −1} (13)

In Expression (13), PT is the start probability PT (i) of the target track, and is expressed as Expression (6) described above.
Further, PF is a false track start probability PF (i), and is represented by the above-described equation (11).

  In the third embodiment of the present invention, the predicted tracking start performance value calculated by the tracking start performance calculation unit 3 and the tracking start performance calculated by the tracking start performance totaling unit 7 are set at a specific time t ( When the start probability PT (i) and PF (i) of the target track and false track in i) are considered to correspond to the detection probability and false alarm probability of the radar 10, respectively, the detection probability and false alarm probability Are assumed to be virtual signal-to-noise ratios VSNRp and VSNR required by the radar 10 to realize the above.

  As described above, by applying the virtual signal-to-noise ratios VSNRp and VSNR by the radar 10 as the performance index for starting tracking, it is possible to set an observation interval at which the target track is high and the erroneous track start probability is low. It becomes possible.

Embodiment 4 FIG.
Although not particularly mentioned in the first to third embodiments, the simulation observation interval candidate setting unit 4 calculates the tracking start performance at the optimum observation interval expected value Tpre set by the analysis observation interval setting unit 2. The pseudo track generated by the pseudo track generation unit 6 based on the pseudo start value calculated by the pseudo observation value generation unit 5 by applying the predicted tracking start performance value calculated by the unit 3 and the optimal observation interval prediction value Tpre Is used to predict the optimum observation interval without resetting the observation interval candidate when it is determined that the two performance differences between the tracking start performance calculated by the tracking start performance counting unit 7 and the tracking start performance are smaller than a predetermined value. The value Tpre may be output to the radar parameter instruction unit 8 as the final optimum observation interval.

A radar apparatus according to Embodiment 4 of the present invention using the simulation observation interval candidate setting unit 4 having the above function will be described below.
The apparatus configuration according to the fourth embodiment of the present invention is the same as that described above (see FIG. 1), and the processing procedure is also as shown in FIG.

  However, in this case, since the contents of the simulation observation interval candidate setting process (step S5) and the simulation result totaling process (step S8) in FIG. 2 are partially different, the following description is limited to only steps S5 and S8. To do.

First, the simulation observation interval candidate setting unit 4 sets observation interval candidates for performing a simulation in step S5.
Note that immediately after the optimum observation interval predicted value (observation interval) Tpre is input from the analysis observation interval setting unit 2 and when step S5 is first executed, one observation interval Tpre is the observation interval for simulation. Be a candidate.

  Here, when the observation interval candidate for simulation is the observation interval Tpre, the tracking start performance value PT (i) obtained in the simulation result totaling process (step S8) at the previous execution and the tracking start performance calculation process (step S3) ) In the simulation interval candidate setting process (step S5) when the performance difference from the tracking start performance Ptgt (i) at the observation interval Tpre calculated in () exceeds a predetermined threshold value ΔP (a parameter set in advance). A simulation is performed for two types of observation intervals: a value increased from the interval Tpre by the step width ΔT and a reduced value.

  On the other hand, the performance difference between the tracking start performance value PT (i) obtained in the previous step S8 and the tracking start performance Ptgt (i) at the observation interval Tpre calculated in step S3 does not exceed the predetermined threshold ΔP. In this case, at this stage, the search of the observation interval by simulation is terminated, and the simulation observation interval candidate setting unit 4 outputs the optimum observation interval expected value Tpre to the radar parameter instruction unit 8 as the optimum observation interval.

  As described above, the simulation observation interval candidate setting unit 4 according to the fourth embodiment of the present invention compares the accuracy of the predicted optimum observation interval Tpre obtained by analysis with the simulation, and if the accuracy is high, the simulation is executed. Since it is configured to cancel, the optimal observation interval can be determined early.

  1 observation condition input unit, 2 observation interval setting unit for analysis, 3 tracking start performance calculation unit, 4 observation interval candidate setting unit for simulation, 5 pseudo observation value generation unit, 6 pseudo track generation unit, 7 tracking start performance calculation unit, 8 Radar parameter indicating unit, 9 Radar control unit, 10 Radar.

Claims (5)

A radar for searching a predetermined observation area;
A radar control unit for controlling the radar;
Tracking start means for detecting an unknown target by tracking the time series data of the observation values by the radar and generating a target track,
In the radar apparatus that periodically provides the observation value from the radar to the tracking start means,
A simulation processing means for setting an observation interval, which is one of the parameters of the observation value, to a value at which an actual tracking start performance in the tracking start means is optimal;
The simulation processing means includes:
An observation interval setting unit for searching for an observation interval in performance analysis and setting an observation interval candidate and generating an optimum observation interval prediction value;
The predicted tracking start performance value based on the observation interval candidate set by the analysis observation interval setting unit is analytically predicted from the radar observation performance value based on specific assumptions relating to wake generation. Tracking start performance calculation unit,
An observation interval candidate setting unit for simulation that searches for observation intervals in the simulation and sets candidate observation intervals,
A pseudo-observation value generator for generating a pseudo-observation value to be used for tracking start simulation based on the observation performance value of the radar;
A pseudo wake generation unit that generates a pseudo wake with a tracking processing function equivalent to a tracking start means in which a radar provides an observation value, using the pseudo observation value generated by the pseudo observation value generation unit;
A tracking start performance totaling unit that calculates and calculates the tracking start performance using the pseudo wake generated by the pseudo wake generation unit,
A radar apparatus comprising: a radar parameter instruction unit that instructs a setting value of an observation interval determined by simulation to the radar control unit. The tracking start performance expected value calculated by the tracking start performance calculation unit and the tracking start performance calculated by the tracking start performance totaling unit are:
The radar apparatus according to claim 1, wherein the radar apparatus has a start probability of a target wake at a specific time. The tracking start performance expected value calculated by the tracking start performance calculation unit and the tracking start performance calculated by the tracking start performance totaling unit are:
The radar apparatus according to claim 1, wherein the radar apparatus has a mistrack start probability at a specific time. The tracking start performance expected value calculated by the tracking start performance calculation unit and the tracking start performance calculated by the tracking start performance totaling unit are:
In order to realize the detection probability and the false alarm probability when the start probability of the target track and the false track at a specific time is considered to correspond to the detection probability and the false alarm probability of the radar, respectively. The radar apparatus according to claim 1, wherein the radar apparatus has a virtual signal-to-noise ratio required by the radar. The simulation observation interval candidate setting unit is
The tracking start performance predicted value calculated by the tracking start performance calculation unit in the optimal observation interval predicted value set by the analysis observation interval setting unit,
Based on the pseudo-observation value generated by the pseudo-observation value generation unit by applying the optimal observation interval predicted value, the tracking start performance totaling unit calculated using the pseudo wake generated by the pseudo-track generation unit Tracking start performance,
If it is determined that the two performance differences are smaller than a predetermined value, the optimum observation interval predicted value is set as the final optimum observation interval in the radar parameter instruction unit without resetting the observation interval candidate. The radar apparatus according to claim 1, wherein the radar apparatus outputs the radar apparatus.

Images