Fig. 20From: Deep learning for inverse design of low-boom supersonic configurationsOptimized near-field signature matching a physically non-existent targetBack to article page