This document presents research on classifying hand movements using surface electromyography (sEMG) signals and a multichannel convolutional neural network. The researchers extracted power spectral density features from sEMG datasets using Burg's method and fed these features into a CNN model with two parallel channels for classification. The model achieved average accuracies of 99.63% and 98.22% on the two datasets. Future work could involve adding more layers to the CNN or extracting different feature sets to improve classification performance, especially on more challenging subsets of the data.