基於前庭感知的人類站立姿勢控制

Abstract

現有的動作生成方法大多缺少參考感知器官資訊的影響。本論文中，我們提出了一個感知運動架構，藉由模擬感知資訊對動作的影響來生成人物在動暈症下的平衡動作。前庭系統、視覺和自體感知提供感知資訊使大腦採取適當的平衡控制。由於前庭系統感知為平衡控制的關鍵，我們的研究中使用前庭系統模型接收人物真實資訊處理成感知資訊，再利用平衡控制器根據取得的感知資訊採取適當的動作命令以達到平衡控制。前庭系統會因為持續的運動而產生動暈症(motion sickness)，導致人會有暈眩且動作不平衡的情況產生。本論文著重於模擬前庭感知對平衡控制的影響。我們還討論了不同刺激大小對動暈症和平衡動作的影響。我們整合感知系統與擬人的平衡控制架構去生成平衡動作。

Existing motion generation methods for character animation rarely consider the influences of the perceived information from sensory organs. In this thesis, we propose a sensorimotor framework to generate balancing motions under motion sickness by simulating the effects of the perceived sensory information to motions. The vestibular, visual, and somatosensory systems provide sensory information to the brain to produce appropriate balance control. Because the vestibular perception is crucial of balance control, our approach achieves balance control by adopting a vestibular system model to transfer real information into perceived information, and then applying the balance controller to generate appropriate action commands based on the acquired perceived information. The vestibular system will produce motion sickness due to continuous movement, it will lead people to be dizziness and unbalance. We focus on simulating balance control motion with the consideration of the influence of vestibular model. We also discussed the effect of stimulus size on motion sickness and balance motion. We integrate the sensory system and the humanoid balance control structure to generate balance motion.