Principle of Existing GMTI Algorithms
All existing GMTI algorithms use the Doppler characters of moving targets and their differences from those of the stationary targets to accomplish the target detection. There are mainly two parts in GMTI algorithms: clutter suppression and CFAR detection.
Principle of Single-Antenna GMTI Algorithms
Different GMTI algorithms use different methods to suppress the clutter. By suppressing the clutter, the spectra of moving targets can be separated from the clutter, and the SCR is increased.
(a) Spectrum filtering algorithm
According to the analysis in Sect. 2.3.2, a Doppler centroid shift is induced by the cross-track velocity of a moving target. The spectrum of the moving target will move towards the high-band of the PRF with the increase of the cross-track velocity, and the spectrum of the target will be eventually located outside the clutter. Based on the Doppler centroid shift, the moving target can be detected by setting up a series of band-pass filter in the high-band of PRF. Spectrum within the band-pass filter will be restrained, and the clutter will be suppressed. After the clutter suppression, the spectrum of the moving target can be extracted, and the target can be re-focused, as shown in Fig. 2.6.
In Fig. 2.6, n band-pass filters with different frequency center are set in the range-Doppler domain, and the clutter is not located within the band-pass filters. If the target has only the cross-track velocity, the target can be focused after shifting its spectrum into baseband. The flowchart of spectrum filtering algorithm is illustrated in Fig. 2.7.
Fig. 2.6 Principle of the spectrum filtering algorithm
Fig. 2.7 Flowchart of the spectrum filtering algorithm
However, the spectrum filtering algorithm requires that the cross-track velocity of the moving target will not induce significant RCM, i.e., the RCM of the target must be less than half of the range resolution.
where b is the beam angle, qr denotes the range resolution.
Moreover, the spectrum filtering algorithm requires that there is no Doppler ambiguity. The cross-track velocity has to satisfy the condition that
where i is the coefficient of antenna and signal processing.
The Doppler ambiguity will be induced in most practical applications. However, although the spectrum filtering algorithm requires that no Doppler ambiguity exists, it can still suppress the clutter only if the spectrum of the moving target locates outside the clutter.
In conclusion, there are mainly three limitations of spectrum filtering algorithm: first, the velocity detection range has a inconsistent requirement with the observation swath; second, the spectrum filtering algorithm cannot detect targets with slow cross-track velocity and blind velocity; third, targets with only along-track velocities cannot be detected.