Reactive Anticipatory Robot Skills with Memory
| Type of publication: | Book chapter |
| Citation: | Girgin_RR_2023 |
| Booktitle: | Robotic Research |
| Year: | 2023 |
| Pages: | 436-451 |
| Publisher: | Springer |
| Crossref: | Girgin_ISRR_2022: |
| Abstract: | Optimal control in robotics has been increasingly popular in recent years and has been applied in many applications involving complex dynamical systems. Closed-loop control strategies include model predictive control (MPC) and time-varying linear controllers optimized through iLQR. However, such feedback controllers rely on the information of the current state, limiting the range of robotic applications where the robot needs to remember what it has done before to act and plan accordingly. The recently proposed system level synthesis (SLS) framework circumvents this limitation via a richer controller structure with memory which gets feedback on the history of the state. In this work, we propose to design reactive anticipatory robot skills with memory via the SLS framework by extending it to tracking problems involving nonlinear systems and nonquadratic cost functions, which greatly extends the domain of applications of the standard SLS framework. We showcase our methods with a pick-and-place task on a 7-axis Franka Emika robot and a bimanual handover task in a simulated environment. |
| Keywords: | |
| Projects |
Idiap |
| Authors | |
| Editors | |
| Added by: | [UNK] |
| Total mark: | 0 |
|
Attachments
|
|
|
|
|
Notes
|
|
|
|
|
