An innovative array diagnosis technique based on a compressive-sensing (CS) paradigm is introduced in the case of linear arrangements. Besides detecting the faulty elements, the approach is able to provide the degree of reliability of such an estimation. Starting from the measured samples of the far-field pattern, the array diagnosis problem is formulated in a Bayesian framework and it is successively solved with a fast relevance vector machine (RVM). The arising Bayesian compressive sensing (BCS) approach is numerically validated through a set of representative examples aimed at providing suitable user's guidelines as well as some insights on the method features and potentialities.
11 Figures and Tables
Fig. 1. Illustrative Example (Taylor, , , ,
Fig. 2. Sensitivity Analysis (Taylor, , , ,
Fig. 3. BCS Assessment ( , )–Behaviour of and versus SNR for (a) Taylor array and (b) Dolph array. Diagnosis of a
Fig. 4. Comparative Assessment ( , —Behaviour of versus SNR: (a) Taylor array and (b) Dolph array.
Fig. 5. Comparative Assessment (Dolph, )—Behaviour of versus SNR obtained by the BCS and the approaches when (a) ,
Fig. 6. Comparative Assessment [ , , ,
Fig. 7. Non-Regular Sampling (Taylor, , , ,
Fig. 8. Non-Regular Sampling (Taylor, , , ).
Fig. 9. Non-Regular Sampling (Dolph, , ,
Fig. 10. Far-Field vs. Near-Field Sampling (Dolph, ,
Fig. 11. Detection of partial failures (Dolph, , ,
Download Full PDF Version (Non-Commercial Use)