The second branch guided by the orange arrow implements the estimation of the DOA for each source; the activation function of this branch is the Leaky ReLU function. This branch is a regression problem; its loss function is the mean squared difference loss function, whose expression is: 𝐿(...
Therefore, the ReLU function was adopted as the activation function here. The training of the auto-encoder was performed end-to-end. The weights of the auto-encoder were generated simultaneously. The trained weights 𝑤𝑒we of the auto-encoder could be utilized as inputs for path-planning ...