Improved Generative Adversarial Network with Hybrid Attention Mechanism for Path Planning
Yifan Zhang, Tao Sun, Song Ma · 2023
Multi-agent path planning is one of the most challenging problems in the field of artificial intelligence. In this paper, an improved generative adversarial network model with hybrid attention mechanism for data-driven path planning is proposed. Firstly, by introducing spatial attention mechanism, positional attention mechanism and channel attention mechanism, a hybrid attention module is designed. This module not only establishes feature relationship between different locations, but also provides adaptive feature weights to highlight key information. Next, by embedding the hybrid attention module into the generative adversarial network, an improved generative adversarial network model is developed. The proposed model can capture effective feature information in the image and ensures that the generated path can evade obstacles accurately. Finally, some simulation results are provided to verify the superiority and accuracy of the model in path planning.