Radar antenna side lobe identification algorithm based on signal magnitude-phase characteristics

Abstract

The invention provides a radar antenna side lobe identification algorithm based on signal magnitude-phase characteristics. Sum and difference signals are classified according to different magnitude-phase characteristics of the radar sum and difference signals at a main lobe and a side lobe of a radar antenna by adopting a support vector machine classification method, and then the main lobe and the side lobe are identified, so that the problem of side lobe identification in radar measurement is solved.

Description

Technical field

[0001] The invention is used for the research of the antenna sidelobe recognition problem in radar measurement. Background technique

[0002] In radar measurement, the lobes in the direction of maximum emission power are the main lobe, and the rest are side lobes or side lobes. When the main lobe of the antenna receives a signal, the signal gain is the largest, the signal-to-noise ratio is the highest, and the operating distance is the farthest. The antenna captures and tracks the target by using the main lobe. However, because the antenna's side lobes also have several convergent tracking points, when the antenna's electrical axis is not aligned with the target and the signal is strong, the side lobes may also capture the target and form tracking, causing false images. Since the gain of the first side lobe is relatively high and it is closer to the main lobe, when the tracking target is closer, the signal is strong, and the ground equipment receiving level is h...

Examples

Embodiment Construction

[0023] Steps 1) to 4) are implemented in sequence according to the algorithm of the present invention.

[0024] Step 3 of the algorithm of the present invention is to use support vector machines to classify and train the normalized eigenvectors, wherein the training sample collection method is preferably to collect experimental samples according to the ballistic guidance method of the spiral curve, and this sample selection method can guarantee the training samples The comprehensiveness, diversity and correctness of the training sample classification. Among them, the support vector machine is preferably the SVMPerf classification tool, which has high training efficiency.

[0025] In the algorithm step 4) of the present invention, the constraint relationship between the values ​​of the main lobe confidence threshold α and the side lobe confidence threshold β is α≥β, -2≤α≤2, -2≤β≤2. When the main lobe confidence threshold α is -0.13 and the side lobe confidence threshold β is -0.13, ...

Technical field

[0001] The invention is used for the research of the antenna sidelobe recognition problem in radar measurement. Background technique

[0002] In radar measurement, the lobes in the direction of maximum emission power are the main lobe, and the rest are side lobes or side lobes. When the main lobe of the antenna receives a signal, the signal gain is the largest, the signal-to-noise ratio is the highest, and the operating distance is the farthest. The antenna captures and tracks the target by using the main lobe. However, because the antenna's side lobes also have several convergent tracking points, when the antenna's electrical axis is not aligned with the target and the signal is strong, the side lobes may also capture the target and form tracking, causing false images. Since the gain of the first side lobe is relatively high and it is closer to the main lobe, when the tracking target is closer, the signal is strong, and the ground equipment receiving level is h...

Examples

Embodiment Construction

[0023] Steps 1) to 4) are implemented in sequence according to the algorithm of the present invention.

[0024] Step 3 of the algorithm of the present invention is to use support vector machines to classify and train the normalized eigenvectors, wherein the training sample collection method is preferably to collect experimental samples according to the ballistic guidance method of the spiral curve, and this sample selection method can guarantee the training samples The comprehensiveness, diversity and correctness of the training sample classification. Among them, the support vector machine is preferably the SVMPerf classification tool, which has high training efficiency.

[0025] In the algorithm step 4) of the present invention, the constraint relationship between the values ​​of the main lobe confidence threshold α and the side lobe confidence threshold β is α≥β, -2≤α≤2, -2≤β≤2. When the main lobe confidence threshold α is -0.13 and the side lobe confidence threshold β is -0.13, ...

Images