Multi-Objective Optimization of Path Planning based on WCNN and Channel Attention Mechanism

Conventional methods for robot motion control and path planning often lack the reactive and flexible characteristics found in natural systems. Inspired by the brain's navigation mechanism using place cells-neurons that cognitively represent specific environments-a multi-objective optimization model to enhance computational efficiency and accuracy in path planning based on the Width-extended Convolutional Neural Network (WCNN) and Channel Attention Mechanism(CAM) is presented. Specifically, the width-extended convolutional neural network is expanded to intensify feature extraction within convolutional layers, without increasing network depth, thereby capturing crucial path information. Subsequently, channel attention mechanism prioritizes significant path features and constraints to enhance planning precision and minimize errors. Experimental validation on large-scale path planning datasets demonstrates that the method surpasses state-of-the-art multi-objective optimization of path planning techniques by achieving an optimal balance between accuracy and adaptability in path planning.