Inertia Partitioning Modular Robust Control Framework for Reconfigurable Multibody Systems

Abstract

A novel modular modeling and control framework based on Lagrangian mechanics is proposed for multibody systems, motivated by the challenges of modular control of systems with closed kinematic chains and by the need for a modeling framework that remains locally updatable under reconfiguration of body-level geometric and inertial properties. In the framework, modularity is defined with respect to the degrees of freedom of the multibody system, represented in the model by the minimal generalized coordinates, and the inertial properties of each body are partitioned with respect to how they are reflected in the kinetic energy of the system through the motion induced by each degree of freedom. By expressing body contributions through body-fixed-frame Jacobians and spatial inertia matrices, the dynamic model remains locally updatable under changes in geometric and inertial parameters, which is advantageous for reconfigurable multibody systems. For multibody systems in which a mapping between the auxiliary and minimal generalized coordinates is available, the approach accommodates closed kinematic chains in a minimal-coordinate ordinary-differential-equation form without explicit constraint-force calculation or differential-algebraic-equation formulation. Based on the resulting modular equations of motion, a robust model-based controller is designed for trajectory tracking, and practical boundedness of the tracking error is analyzed under bounded uncertainty and external disturbance. The proposed framework is implemented in simulation on a three-degree-of-freedom series-parallel manipulator, where uncertainties and disturbances are introduced to assess robustness. The results are consistent with the expected stability and tracking performance, indicating the potential of the framework for trajectory-tracking control of reconfigurable multibody systems with closed kinematic chains.