Research on Robot Path Optimization Algorithm Based on LSTM-Improved Ant Colony Algorithm

In the field of computer-aided robotics, trajectory scheduling is crucial for multi-robot systems operating in complex, dynamic, and different kinds of environments. This paper introduces a novel trajectory optimization algorithm that integrates the Long Short-Term Memory (LSTM) network with the traditional Ant Colony Algorithm (ACA) to address the limitations within the conventional ACA, such as slow convergence and susceptibility to local optima. The incorporation of the LSTM module enhances the ants' ability to predict environmental changes, improving the overall algorithm performance, especially regarding energy efficiency. This proposed approach specifically tackles the often-overlooked issue of optimizing energy consumption in existing trajectory scheduling algorithms. Experimental results validate the effectiveness of the LSTM-enhanced ACA, showing substantial improvements in both trajectory length reduction and energy efficiency over traditional methods. Indoor environment tests reveal that the algorithm increased the search efficiency by 45.2%, reduced the trajectory length by 11.2%, and decreased the number of the optimal trajectory turning points by 42.9%. These findings suggest that the LSTM-enhanced ACA not only boosts trajectory scheduling accuracy but also optimizes energy usage, making it particularly well-suited for practical applications in multi-robot systems. This research presents a robust solution for energy-efficient trajectory scheduling, offering significant potential for advancing autonomous robotics.