The clutter and noise correlation matrix estimated from system received data almost comprises the entire information about clutter and noise.An eigenvalue division(EVD) of the matrix can provide an eigenspace and within the space one can rank the power of the clutter and noise according to the metrics of eigenvalues whose eigenvectors are statistically independent and orthogonal to each other.Studies on adaptive processing especially on reduced-rank processing based on eigenspace can help probing the optimum approach in theory.The cross spectral method(CSM)

minimum power eigencanceler(MPE) and minimum norm eigencanceler(MNE) are discussed respectively in this paper and the intrinsic relationship among them from the view of eigenstructure is revealed as well.Finally

the clutter suppression performance is compared by means of simulation.