Optimal Parameter Identification for Discrete Mechanical Systems with Application to Flexible Object Manipulation

摘要

We present a method for system identification of flexible objects bymeasuring forces and displacement during interaction with a manipulating arm.We model the object's structure and flexibility by a chain of rigid bodiesconnected by torsional springs. Unlike previous work, the proposed optimalcontrol approach using variational integrators allows identification of closedloops, which include the robot arm itself. This allows using the resultingmodels for planning in configuration space of the robot. In order to solve theresulting problem efficiently, we develop a novel method for fast discrete-timeadjoint-based gradient calculation. The feasibility of the approach isdemonstrated using full physics simulation in trep and using data recorded froma 7-DOF series elastic robot arm.