Hybrid Beamforming System Diagnosis: Failure Modeling and Identification

With the rapidly increasing scale and complexity, millimeter-wave communication systems are getting more powerful, but are also less reliable. Due to hardware characteristics and connecting structures, millimeter-wave communication systems may experience device faults which result in performance degradation or even failure. Therefore, system diagnosis techniques for fault detection and localization are essential to system monitoring and maintenance. In this paper, two compressed sensing based diagnosis techniques are proposed, which can jointly detect and locate faulty antennas, phase shifters, and RF chains. The first technique is applicable to the diagnosis in a multipath-free environment where the channel state information can be obtained based on spatial coordinates. Furthermore, we propose the second technique that does not require the channel state information, which is suitable for outdoor online diagnosis in a multipath scattering environment. Finally, numerical simulations show that the proposed techniques can jointly detect all faulty devices with high probability.