www.radartutorial.eu www.radartutorial.eu Radar Basics

Clutter-Doppler Filter (2)

Another method of clutter suppression is known as the “Moving Window” method¹). In this method, the echo data from two complete pulse repetition periods (PRP) are initially stored without weighting. Only during processing are these data and the echo data from the current pulse repetition period multiplied by weighting coefficients and subsequently added. This method is frequently used to suppress echoes from stationary targets. It is particularly suitable for software implementation because the originally stored data remain unchanged and are therefore available for further processing.

I&Q Data
Memory
Memory
Multiplier
Multiplier
Multiplier
Adder
Rescaler
(3rd PRT)
(2nd PRT)
(1st PRT)
Weights
Weights
Weights

Figure 1: Block diagram of a Clutter-Doppler Filter using the “Moving Window” method

I&Q Data
Memory
Memory
Multiplier
Multiplier
Multiplier
Adder
Rescaler
(3rd PRT)
(2nd PRT)
(1st PRT)
Weights
Weights
Weights

Figure 1: Block diagram of a Clutter-Doppler Filter using the “Moving Window” method

I&Q Data
Memory
Memory
Multiplier
Multiplier
Multiplier
Adder
Rescaler
(3rd PRT)
(2nd PRT)
(1st PRT)
Weights
Weights
Weights
I&Q Data Multiplier Multiplier Multiplier Memory Memory Adder Rescaler

Figure 1: Block diagram of a Clutter-Doppler Filter using the “Moving Window” method (interactive diagram)

The echo data are weighted in the multiplier stages only when the third pulse repetition period is processed. The weighting coefficients can be selected in the same way as for the “Integrate & Dump” method. For example, the echo data from the first and third pulse repetition periods are each multiplied by +0.5, while the data from the second, middle period are multiplied by −1. Echo signals with identical amplitudes and phases therefore cancel each other during the subsequent addition. However, if the phase of the echo signal changes between pulse repetition periods, a nonzero output signal is generally produced.

In this method, arithmetic operations and possible carries may also increase the number of bits required to represent the result. A rescaler must therefore adjust the result to the required bit width. Often, the least significant bits are simply discarded. However, this may cause very weak echo signals to be lost. More sophisticated rescaling methods are therefore also used.

The main difference compared with the “Integrate & Dump” method is that the latter produces only one output signal from a group of pulse repetition periods used for MTI processing. In contrast, the “Moving Window” method uses three consecutive pulse repetition periods for each output signal. With each additional pulse repetition period, the processing window advances by one position, allowing a new output signal to be calculated. If a total of N pulse repetition periods are used for MTI processing, (N−2) output signals are produced. This is particularly advantageous when a radar requires a larger number of echo signals for pulse integration. For example, this may be important in radar systems that do not employ monopulse techniques.

Description of the Modules in the Block Diagram

¹) Do not confuse this method with the Sliding Window used in the plot extractor!