Identifiability analysis for array shape self‐calibration in colocated multiple‐input multiple‐output radar using Cramér–Rao bound

In this study, parameter identifiability in array shape self‐calibration in colocated multiple‐input multiple‐output radar is addressed under a deterministic framework. In contrast to the random model used in the previous analysis, some distinct identifiability conditions are established through deriving and then analysing the Cramér–Rao bound on self‐calibration accuracy of antenna positions using far‐field targets whose directions of arrival and scattering coefficients are initially unknown. It is proved that at least three non‐collinear targets are needed to precisely self‐calibrate the positions of antennas of arbitrary geometry when there exist a position reference and a direction reference . The sole exception is an actually linear array for which self‐calibration is impossible.