JPH0523172U - Tracking filter device - Google Patents

Abstract

(57) [Abstract] [Purpose] The purpose of the tracking filter of a radar device is to obtain a stable tracking filter effect against temporary fluctuations in tracking distance and lack thereof. [Configuration] A fluctuation of a tracking distance equal to or more than a prescribed value is determined by a low-pass filter 9, and a tracking filter freeze processing unit 7 temporarily fixes a filter gain without passing through a prediction calculation unit 5 and a smoothing calculation unit 6 to perform processing. To continue. Further, the low-pass filter with missing pulse determination 10 determines a missing pulse, performs pulse interpolation, and continues processing.

Description

[Detailed description of the device]

[0001]

[Industrial applications]

 The present invention relates to a method of performing prediction and smoothing by a tracking filter of a tracking radar device, and relates to a method of obtaining a stable filter effect even for protruding information.

[0002]

[Prior Art]

 This type of tracking filter method expands the tracking target distance for each pulse in time series obtained from the tracking radar device into a filter state model and adjusts the filter gain given to the state vector for each pulse. However, it predicts and smooths future states based on past states.

As a method of adjusting the filter gain of the tracking filter for each pulse, there is a Kalman filter or the like.

This is an effective method when the target object is performing a stable motion based on a certain equation of motion.

However, when the distance to the tracking target obtained from the tracking radar device fluctuates greatly due to external factors such as noise, or when a series of time-series pulses are lost even at the moment of blinking. In this case, since the distance information that is significantly different from the true value is processed, the filter gain that was asymptotically converged becomes unstable, and even if a stable value is input later, prediction is performed. The smoothing effect has a drawback that it is significantly reduced and the filter gain diverges.

For example, FIG. 5 is a diagram illustrating a Kalman filter of a conventional tracking filter. In the figure, 1 is a state vector in which a target distance is expanded, 2 is a prediction vector, 3 is a smooth vector, 4 is a feedback smooth vector, 5 is a prediction calculation unit, and 6 is a smooth calculation unit.

FIG. 6 is a diagram for explaining a tracking defect of the conventional tracking filter. In the figure, 12 is the actually input distance information, and 13 is the distance after the tracking filter.

First, the operation of FIG. 5 will be described.

The target distance obtained from the tracking radar device is expanded into a state vector 1 of arbitrary dimension,

The prediction vector 2 is calculated by inputting both of the smoothing vector 4 of the target distance of the pulse one pulse before the present into the prediction calculating section 5 and inputting to the smoothing calculating section 6 to adjust the filter gain. To obtain a smooth vector 3. This smooth vector 3 is fed back to the prediction calculation unit 5 and becomes a feedback smooth vector 4 which is synthesized by the next pulse.

Next, the drawbacks of this tracking filter will be described. When there is no large variation in the target distance obtained for each pulse, the filter gain calculated by the smoothing calculation unit 6 by approaching the loop described above approaches the asymptote and moves toward the convergence.

However, when the target distance of the n-th pulse in FIG. 6 temporarily changes greatly like the distance information 12 actually input due to noise or the like, the smoothing calculation unit 6 detects the information of this protrusion. Even if the feedback smoothing vector 4 changes greatly to smooth using the filter gain of the immediately preceding pulse, and the target distance of the next pulse is not more prominent than the current one as in the post-tracking filter distance 13 in FIG. The effect is delayed and propagates, disturbing the filter gain that was converging toward the asymptote and moving away from the asymptote, or a stable smooth vector can be obtained because the filter gain diverges. There were no drawbacks.

[0013]

[Problems to be Solved by the Invention] Since the conventional tracking filter method is configured as described above, the filter gain is large when a target distance that is protruding even once is input during a series of time series pulses. Because of fluctuations, there are problems such as divergence and the inability to perform accurate prediction and smoothing processing until the filter gain stabilizes.

The present invention has been made to solve the above problems, and an object thereof is to obtain a tracking filter method capable of performing stable and accurate prediction and smoothing processing even when the target distance suddenly changes. And

[0015]

[Means for Solving the Problems]

 The tracking filter method according to the present invention suppresses the disturbance of the filter gain by using the low-pass filter that gives the target change allowable range and suppresses the interpolation and function even when the temporarily projected target distance is input. It is added to enable stable prediction and smoothing.

[0016]

[Action]

 By adding a low-pass filter, the allowable range of change of the target distance was limited, the disturbance of the filter gain was suppressed, and stable prediction and smoothing processing were made possible.

Further, the reliability is improved by allowing the processing to be continued on the same principle with respect to data loss.

[0018]

【Example】

 Example 1. An embodiment of the present invention will be described below with reference to the drawings. In FIG. 1, 7 is a tracking filter freeze processing unit, 8 is a frozen smooth vector, and 9 is a low-pass filter. FIG. 2 is a diagram for explaining the principle of one embodiment of the present invention. If the target distance does not change significantly and the state vector 1 is within a certain specified value, the state vector 1 that has passed through the low-pass filter 9 is input to the prediction calculation unit, and the conventional tracking filter described in FIG. Do the same.

As shown in FIG. 2, when the distance change of the n-th pulse exceeds the specified value ΔR, the low-pass filter 9 determines that the tracking filter is unqualified, and the smoothing calculation unit 6 of the tracking filter 6 The filter gain calculated at is fixed as it is calculated at the (n-1) th pulse, and the state vector 1 is input to the tracking filter freeze processing unit 7 instead.

At this time, the tracking filter freeze processing unit 7 stores the prediction vector 2 of the (n-1) th pulse calculated by the prediction calculation unit 5, and outputs this value as the nth smooth vector 8. To do.

Example 2. In the above embodiment, the low-pass filter 10 is provided to control the variation amount of the state vector 1. However, the low-pass filter 11 may be provided inside the low-pass filter 9 with a missing judgment added. FIG. 3 is a processing block diagram of this embodiment. FIG. 4 is a diagram for explaining the principle of this embodiment.

As shown in FIG. 4, when the n-th pulse is missing and a series of time series data is missing due to some influence, the low-pass filter with missing pulse determination 10 activates the freeze processing unit with interpolation function 11. Then, the prediction vector 2 calculated in the n-1st pulse is input as the state vector 1 of the nth pulse while maintaining the filter gain calculated in the n-1th pulse, and the smooth vector 3 is input. calculate.

[0023]

[Effect of the device]

 As described above, according to the tracking filter of the present invention, since the low-pass filter is provided in the front part of the tracking filter, a stable tracking filter effect can be obtained without being affected by a large change in the state vector.

Further, by providing the missing pulse determination unit inside the low-pass filter, it is possible to obtain the tracking filter effect without stopping the processing even for missing information.

[Brief description of drawings]

FIG. 1 is a functional block diagram of an embodiment of the present invention.

FIG. 2 is a diagram for explaining the principle of an embodiment of the present invention.

FIG. 3 is a functional block diagram of another embodiment of the present invention.

FIG. 4 is a diagram for explaining the principle of another embodiment of the present invention.

FIG. 5 is a functional block diagram of a conventional tracking filter.

FIG. 6 is a diagram for explaining the principle of a defect of a conventional tracking filter.

[Explanation of symbols]

 1 state vector 2 prediction vector 3 smoothing vector 4 feedback smoothing vector 5 prediction calculation unit 6 smoothing calculation unit 7 tracking filter freeze processing unit 8 frozen smooth vector 9 low pass filter 10 low pass filter with missing pulse judgment 11 freeze processing unit with interpolation function 12 actual Distance information input to 13 Distance after tracking filter

Claims (1)

[Scope of utility model registration request] 1. In a tracking radar device, a tracking change range that a tracking filter can accept is given in advance to a tracking filter that predicts and smooths a tracking distance of a target object (hereinafter referred to as a low-pass filter in a narrow sense). A tracking filter device characterized in that, when a tracking distance that exceeds temporarily is input, the distance information is rejected and the processing is continued.