Efficient Calibration of Colocated MIMO Radar

Measurement-based calibration with a complete MIMO radar system is beneficial due to all imperfections being considered simultaneously. However, if high performance is needed, the number of required calibration measurements escalates with an increasing number of antenna elements. This paper presents a calibration technique that reduces the number of measurements required to obtain adequate calibration coefficients, compensating for all system imperfections that can be described in matrix form. This includes channel imbalance and mutual coupling under the assumption of minimum scattering antennas. The proposed technique estimates the calibration matrix for the transmit and receive array separately, by canceling out the element radiation pattern and effects from the other array through normalization. In this way, the computational complexity and the risk of local convergence is reduced. Using prior knowledge of the expected mutual coupling, the number of equations in the nonlinear least-squares can be further reduced without affecting the calibration performance. Besides validation with a simulated model, the calibration technique is applied to an eight-by-eight MIMO radar system.