Advances in Humanoid Robot Walking Technologies: A Review

摘要

Humanoid robots are regarded as versatile robot platforms that are capable of traversing diverse terrains and performing dangerous or challenging tasks on behalf of humans. Particularly, the development of robust bipedal walking capabilities is deemed crucial for these robots to effectively operate in real-world environments. Despite significant advancements in bipedal walking algorithms over the past three decades, much work is still needed for humanoids to walk stably in diverse environments under disturbances. This paper explores the trends in humanoid walking research spanning the last three decades. Specifically, it primarily categorizes the research trends into model-based and learning-based walking algorithms. The research within each classification of walking algorithms is introduced broadly. Finally, the paper outlines potential directions for future advancements in this field.