Airborne radar adaptive beamforming algorithm based on convex optimization learning

The invention discloses an airborne radar adaptive beamforming algorithm based on convex optimization learning. The airborne radar adaptive beamforming algorithm comprises the following steps of: step1, area array receiving signal modeling; 2, autocorrelation matrix estimation, and iteration step length determining; 3, gradient iteration outer loop starting; step 4, random gradient iteration internal circulation starting; step 5, final vector output during the last time of external circulation; and step 6, adaptive beam forming. The algorithm provided by the invention is based on the above principles and an iteration mode, and can be suitable for the situations of non-differentiable target functions and non-stationary signals. In engineering practical application, for example, an array structure is huge, subarray partitioning can be carried out on the area array, the array output of each subarray is calculated by adopting an SVRGD algorithm, and then a directional diagram of the wholearea array is synthesized