🤖 AI Summary
This paper addresses the fundamental limitation of conventional Euclidean-space impedance control in robotic manipulators operating within unknown environments: its inability to preserve the intrinsic SE(3) kinematic and dynamic geometric structure. To resolve this, we propose the first intrinsic geometric impedance model formulated directly on the SE(3) Lie group. Leveraging Lie group differential geometry and tangent-space modeling, the approach employs an affine connection to characterize dynamics and integrates nonlinear feedback linearization, thereby achieving pose-position coupling, coordinate independence, and singularity-free control. Simulation and real-world robotic arm experiments demonstrate a 40% improvement in contact transition smoothness and a 32% reduction in trajectory tracking error. The method significantly enhances robustness and precision in complex tasks such as curved-surface grinding and fine assembly.
