The technology of autonomous vehicles is flourishing, and safe driving of autonomous vehicles on land is a concern for everyone. Many scholars are researching optimization planning algorithms to improve the safety of autonomous vehicles. This article provides new insights into path planning for autonomous vehicles by adding self attention mechanism to the CNN-LSTM deep neural network, which is an improved LSTM network model. The addition of self attention mechanism increased the timeliness of CNN-LSTM, improving the efficiency and safety of autonomous vehicles from the set starting point to the endpoint. Finally, the effectiveness of the algorithm was verified through simulated map data, and the results showed that the path planning accuracy and computational efficiency of the model were significantly improved. Using a trained network with text saved for simulation increased the efficiency by 12.23% compared to other models.
CNN-LSTM Unmanned Vehicle Path Planning Based on Self-Attention Mechanism
18.10.2024
283045 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch