Physics-Informed Model Reduction With Coriolis/Centripetal Force Terms Cancellation in 4PPa-2PaR Parallel Manipulator

The article proposes a physics-informed dynamic model reduction strategy for the 4PPa-2PaR parallel manipulator to overcome the high computational burden associated with its multiple closed-loop subchains. A Lagrangian-based formulation is employed to rigorously prove that the Coriolis and centripetal force terms associated with the proximal parallelograms cancel out due to structural symmetry. Furthermore, spectral matrix norm analysis is employed to demonstrate that the remaining Coriolis/centripetal terms constitute bounded residues with minimal contribution to total actuating forces. Validated by simulation and physical experiments, the simplified model achieves a high prediction accuracy while improving computational efficiency by over 56.9% with negligible accuracy loss of less than 4.3% of RMS of computed forces, making it suitable for real-time control applications.