一般求解二足機器人軀幹補償技術是假設軀幹高度為常數，但這個假設會導致求解有不小的誤差，故本文提出新的求解方式來降低求解的誤差，且進一步設計一個八連桿二足機器人，利用MSC.VisualNastran 4D與Matlab結合做即時平衡控制，並探討幾個可變因子對軀幹補償技術的影響，包括：步行週期、軀幹質量、跨距及抬腳高度，以期提供相關領域人員發展建立二足機器人的資訊。結果顯示，新的求解方式可以讓實際ZMP更接近預設ZMP，且也成功得到各個可變因子對軀幹補償技術所造成的影響。 Traditionally, the control method of biped walking compensated by the trunk motion assumed that the trunk height is constant. But this is an unrealistic assumption. This thesis proposed to remove the constraint and used MSC.VisualNastran 4D and Matlab to discuss the feasibility and the effect on variable factors of trunk compensation. The variable factors include motion period, trunk weight, step length and step height. The results of simulation show that the new solution method can make the actual ZMP closer to the desired ZMP, and successfully obtain the influence on each variable factors of trunk compensation.
