Active Antenna Array Diagnosis from Far-Field Measurements

A procedure based on Bayesian compressive sensing (BCS) is presented for faster and robust diagnosis of failed elements in large planar antenna arrays. The traditional approaches exhibit some drawbacks in effectiveness and reliability in noisy data. From measured samples of the far-field pattern, planar array diagnosis is formulated within the Bayesian framework and solved with a fast relevance vector machine (RVM). A 10GHz 10 by 10 rectangular microstrip patch antenna array that emulates element failure via zero excitation is then considered to test the proposed procedure. BCS approach provides faster diagnosis and robust to additive noise.