Pulse Integration
Pulse integration is a method used to improve the probability of detection of a target by jointly processing the echoes from several transmitted pulses in successive pulse repetition intervals. A single weak target echo may be lost in the noise, whereas combining several echoes makes it easier to distinguish the target from the noise. For this purpose, the echo signals from different pulse repetition intervals are combined in a radar signal processor. Depending on the position of the pulse integrator in the signal processing chain, a distinction is made between:
No Target
Figure 1: Position of a coherent pulse integrator
No Target
Figure 1: Position of a coherent pulse integrator
Filter
(IF Amplifier)
Integration
Detector
Circuit
No Target
Figure 1: Position of a coherent pulse integrator
- coherent pulse integration;
- noncoherent pulse integration.
Coherent Pulse Integration
In coherent pulse integration, the integrator (or signal processor) is placed before the amplitude detector, allowing the phase of the signals to be taken into account during processing. For the echo signals to reinforce one another when added, their phase relationships must be preserved, and any phase changes must be compensated for when necessary. For moving targets, such phase changes may be caused by the Doppler frequency. The signal processor assigns the echo signals from each pulse repetition interval to individual range cells according to their distance, with the size of these cells depending on the range resolution of the radar. The echo signals from N pulse repetition intervals are then added with the correct phase alignment. After summation, amplitude detection and threshold comparison are performed.
For example, with a maximum range of 75 km and a range resolution of 150 m, there are 75,000/150 = 500 range cells. In a simple digital pulse integrator, the echo signals from N pulse repetition intervals are stored for each of these range cells. The oldest values are subsequently replaced by newer data. In this way, the most recent N pulses can always be processed together.
In analog signal processors, integration is performed using special filters (for example, with storage tubes). In digital radar signal processors, processing can be performed, for example, using a Fast Fourier Transform (FFT), which allows different Doppler frequencies to be evaluated simultaneously.
For a simplified explanation of the operating principle, we assume that the noise in successive pulse repetition intervals is not correlated and that the individual noise values have a mean of zero and follow a Gaussian distribution. Even after summing the signals from several pulse repetition intervals, the noise remains Gaussian-distributed. While the correctly phase-aligned echo signals reinforce one another, the random noise increases much more slowly.
Assume that the signal level remains constant from one pulse repetition interval to the next (corresponding to Swerling case 0 and approximately to Swerling cases 1 and 3). With correct phase alignment, the amplitude of the useful signal increases in proportion to the number N of integrated pulses. Ideally, this improves the signal-to-noise ratio by a factor of N. According to the basic radar equation, this improvement also affects the maximum target detection range.
In Swerling cases 2 and 4, however, the echo amplitude fluctuates independently from pulse to pulse. Simple coherent summation therefore cannot achieve the same integration gain as with an echo signal that remains stable throughout the integration interval. In such cases, noncoherent pulse integration may be advantageous.
Noncoherent Pulse Integration
No Target
Figure 2: Position of a noncoherent pulse integrator
No Target
Figure 2: Position of a noncoherent pulse integrator
Filter
Detector
Integration
Circuit
No Target
Figure 2: Position of a noncoherent pulse integrator
The noncoherent pulse integrator is located after the amplitude detector. The term noncoherent means that the phase information of the echo signal has already been lost at this stage of processing. Therefore, the signals can no longer be added with the correct phase alignment. Instead, the amplitudes or powers of the echo signals from several pulse repetition intervals are combined before threshold comparison. The achievable integration gain is generally lower than with ideal coherent integration.
In older radar systems, a similar effect was achieved through the persistence of the display combined with the persistence of human vision. Several successive weak echoes therefore appeared as a more clearly visible target indication.
Another possibility is a so-called m-out-of-n detector. First, the circuit checks whether the amplitude of each individual pulse exceeds a specified threshold. It then determines whether a target has been detected at least m times during n pulse repetition intervals. For example, in a 3-out-of-5 detector, at least three of five pulses must exceed the threshold before a target detection is reported. Since this method combines individual decisions that have already been made, it is also called binary integration. This method is also known as a dual-threshold detector, because one threshold is applied to the individual pulses and another to the number of positive decisions.
In digital radar signal processing, noncoherent integration can also be performed by a summing circuit that adds the amplitudes or powers of several echo signals. Older analog radar systems often used a simple low-pass filter at this point.