Learning Null Space Projections in Operational Space Formulation
This addresses a gap in learning constraints for robotic or biomechanical systems, though it appears incremental as it builds on existing operational space formulation tools.
The paper tackles the problem of learning null space projection matrices for kinematically constrained systems without prior knowledge of the policy, geometry, or constraint dimensionality, and demonstrates effectiveness across varying dimensionalities and non-linearities.
In recent years, a number of tools have become available that recover the underlying control policy from constrained movements. However, few have explicitly considered learning the constraints of the motion and ways to cope with unknown environment. In this paper, we consider learning the null space projection matrix of a kinematically constrained system in the absence of any prior knowledge either on the underlying policy, the geometry, or dimensionality of the constraints. Our evaluations have demonstrated the effectiveness of the proposed approach on problems of differing dimensionality, and with different degrees of non-linearity.