Optimization of robotic path planning and navigation point configuration based on convolutional neural networks

Area of Science:

• Robotics and Artificial Intelligence
• Computer Vision

Background:

• Traditional robotic path planning algorithms struggle with efficient and accurate area coverage.
• Intelligent algorithms like genetic algorithms and particle swarm optimization show limitations in point layout optimization.

Purpose of the Study:

• To introduce a novel convolutional neural network (CNN)-based approach for optimizing point configuration in robotic path planning.
• To enhance the speed, accuracy, and efficiency of robotic navigation and area coverage.

Main Methods:

• Developed a CNN-based optimization model integrating polygon image features and Gaussian distribution variability.
• Trained the CNN model using datasets from systematic point configurations.
• Defined a coverage index to evaluate path planning effectiveness.

Main Results:

• The proposed CNN model significantly improves robotic navigation and area coverage precision.
• Achieved an experimental error rate of less than 8% on the test dataset.
• Demonstrated superior performance compared to traditional traversal and intelligent algorithms.

Conclusions:

• The CNN-based optimization model is effective for efficient and accurate robotic path planning.
• This approach offers a significant advancement in robotic navigation capabilities.
• The method provides a robust solution for precise area coverage tasks.