Blind Array Calibration of Mutual Coupling, Phase, and Gain for Automotive Radar

Solomon Goldgraber Casspi, Joseph Tabrikian*, Hagit Messer

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review


One of the limiting factors in the performance of radar systems is the presence of mutual coupling (MC) between receive antenna elements or array imperfections, such as antenna phase and gain errors. Therefore, the data model is misspecified, resulting in high sidelobe levels in the beam pattern, low angular resolution, and biased angle estimation. In this article, we propose a blind calibration scheme for uniform planar arrays. Our method is based on multiple measurements of various scenarios, with an arbitrary and unknown number of targets-of-opportunity, unknown directions-of-arrival (DOAs), and unknown intensities. The proposed method is based on spatial smoothing and forward-backward averaging techniques, in order to identify the signal and noise subspaces. In the presence of MC or array imperfections, the signal subspace leaks into the noise subspace. The proposed method seeks to find and compensate for model misspecification using a model-order selection criterion. We evaluate the performance of our method through simulations, in terms of DOA estimation accuracy and resolution. Our results demonstrate that the DOA estimation performance after calibration with our proposed method is close to that of a perfectly calibrated array.

Original languageEnglish
Pages (from-to)1060-1073
Number of pages14
JournalIEEE Transactions on Aerospace and Electronic Systems
Issue number1
StatePublished - 1 Feb 2024


FundersFunder number
Israel Science Foundation2666/19


    • Array calibration
    • automotive radar
    • forward-backward averaging (FBA)
    • model order selection
    • mutual coupling (MC)
    • phase and gain calibration
    • spatial smoothing (SS)
    • target enumeration


    Dive into the research topics of 'Blind Array Calibration of Mutual Coupling, Phase, and Gain for Automotive Radar'. Together they form a unique fingerprint.

    Cite this