Enhanced Operational Space Formulation for Multiple Tasks by Using Time-Delay Estimation

Abstract

The operational space formulation (OSF) has been enhanced from a practical viewpoint through the application of the time-delay estimation (TDE). In principle, the OSF enables an excellent decentralized control owing to its ability to achieve dynamic consistency. In reality, however, it can suffer from modeling errors and relatively large computational demands. As a remedy for these problems, the OSF has been combined with the TDE, which is known for its accurate estimation of robot dynamics with high computational efficiency. By virtue of the TDE, the OSF with the TDE (OSFTDE) shows enhanced accuracy in terms of dynamic consistency and control performance along with enhanced computational efficiency. Through simple but obvious simulations and experiments, the OSFTDE shows much better accuracy than the OSF with modeling error of 5% or higher, although accuracy is slightly worse than the OSF with a perfect model, thereby demonstrating its practical advantages.