“…Baldoni et al [ 175 ] utilize neural networks to guide the generation of path planning with the expectation of improving the generation of datasets and increasing the efficiency of path planning. The authors conclude that the path planning task for narrow passages and maze-like maps is very similar to the image segmentation task, so they choose U-net, which is widely used in the field of image segmentation, to train the path planning task, and the results show that the neural network-guided RRT significantly outperforms the traditional RRT in terms of path planning.…”