• JADCE is formulated as a joint sparse signal recovery problem exploiting the block-type row-sparse structure of mmWave channels in massive MIMO-OFDM systems.
• A difference-of-convex function algorithm (DCA) with multiple measurement vector (MMV) frameworks is proposed to promote row-sparsity and enhance detection/estimation accuracy in strongly coherent systems.
• A fast DCA-based algorithm incorporating a proximal operator and ADMM significantly reduces computational complexity while maintaining performance.
• The proposed DC algorithms outperform state-of-the-art compressed sensing based JADCE techniques in both active user detection and channel estimation accuracy.