Simulation study demonstrates enhanced navigation recall and operation success for electric power equipment, indicating improved adaptability of embodied AI systems.
Key Points
To investigate the application of embodied intelligence based on large-scale pre-trained models for navigation and operational tasks in electric power equipment environments.
Simulated embodied AI agents learning through egocentric perception and direct environmental interaction rather than static datasets.
Evaluated performance across navigation simulations and robotic equipment operation tasks.
Navigation simulation demonstrated a recall enhancement of over 9.74% compared to standard baseline algorithms.
Equipment operation simulations achieved a 50% average task success rate, marking a baseline benchmark in power AI manipulation.