Towards multibody modeling of passive dynamic quadruped robots using MSC ADAMS

Abstract

Passive dynamic walking refers to the motion of legged robots that can walk downhill without any actuation or control, relying solely on gravity. Simulating Passive Dynamic Quadruped Robots (PDQR) in mathematical tools like MATLAB is challenging due to the complex equations of motion involved. While prior research relies on theoretical models, multibody dynamic software like MSC ADAMS remains underexplored for PDQR simulation. This article proposes a framework for multibody modeling of rigid-leg PDQR using MSC ADAMS, addressing this research gap. The MSC ADAMS model was refined using insights from a MATLAB-derived mathematical model, with discrepancies primarily observed during swing-leg impacts. To mitigate these differences, a parameter adjustment procedure was introduced, requiring an initial angular velocity of the stance legs exceeding 100% of their final fixed-point value to achieve stable motion and compensate for initial foot slippage. The framework was further evaluated on slopes of varying steepness (e.g., <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" display="inline" overflow="scroll"> <mml:mn>0.5</mml:mn> <mml:mo>∘</mml:mo> <mml:mo>,</mml:mo> <mml:mspace width="0.25em"/> <mml:mn>1.5</mml:mn> <mml:mo>∘</mml:mo> <mml:mo>,</mml:mo> <mml:mspace width="0.25em"/> <mml:mn>2.5</mml:mn> <mml:mo>∘</mml:mo> </mml:math> ). Results indicate that steeper slopes demand proportionally greater increases in the initial angular velocity of stance legs. Additionally, impact-phase discrepancies become more pronounced on steeper slopes. Overall, the framework demonstrates the efficiency of MSC ADAMS for PDQR modeling, producing results closely aligned with MATLAB's analytical models.