Co-optimizing Physical Reconfiguration Parameters and Controllers for an Origami-inspired Reconfigurable Manipulator

TL;DR

We address the problem of co-optimizing physical reconfiguration parameters and controllers for a tendon-driven origami-inspired manipulator. A forward model based on the minimum potential energy method predicts shape under tendon displacements $D$ and stiffness $S$, while a PPO-based RL framework jointly optimizes the stiffness distribution $oldsymbol{\omega}$ (modeled as a Gaussian $p_\phi(\omega)$) and the control policy $\pi_\theta(a|s,\omega)$ to maximize task return. The study demonstrates that co-optimization enables reaching with obstacle avoidance that is infeasible with fixed reconfiguration, and shows that stiffness choices influence both trajectories and workspace. This framework lays the groundwork for adaptive, post-fabrication robotics that can reconfigure morphology and behavior to tackle diverse tasks and environments.

Abstract

Reconfigurable robots that can change their physical configuration post-fabrication have demonstrate their potential in adapting to different environments or tasks. However, it is challenging to determine how to optimally adjust reconfigurable parameters for a given task, especially when the controller depends on the robot's configuration. In this paper, we address this problem using a tendon-driven reconfigurable manipulator composed of multiple serially connected origami-inspired modules as an example. Under tendon actuation, these modules can achieve different shapes and motions, governed by joint stiffnesses (reconfiguration parameters) and the tendon displacements (control inputs). We leverage recent advances in co-optimization of design and control for robotic system to treat reconfiguration parameters as design variables and optimize them using reinforcement learning techniques. We first establish a forward model based on the minimum potential energy method to predict the shape of the manipulator under tendon actuations. Using the forward model as the environment dynamics, we then co-optimize the control policy (on the tendon displacements) and joint stiffnesses of the modules for goal reaching tasks while ensuring collision avoidance. Through co-optimization, we obtain optimized joint stiffness and the corresponding optimal control policy to enable the manipulator to accomplish the task that would be infeasible with fixed reconfiguration parameters (i.e., fixed joint stiffness). We envision the co-optimization framework can be extended to other reconfigurable robotic systems, enabling them to optimally adapt their configuration and behavior for diverse tasks and environments.

Figures (10)

• Figure 1: Illustration of programmable motion for two serially connected origami-inspired modules chen2022origami. $S_1$ and $S_2$ represent the stiffness of a joint in the top and bottom module, respectively. When $S_1>S1$, the manipulator undergoes motion 1. If $S_1<S_2$, the manipulator undergoes motion 2 with the same actuation.
• Figure 2: A reconfigurable manipulator consisting of two origami-inspired modules connected in series
• Figure 3: Geometry of the deformed module
• Figure 4: The manipulator under fixed actuation sequence exhibits different motions if the VSJs have different stiffnesses. Trajectories 1, 2, and 3 correspond to stiffness set $S_1$, $S_2$, $S_3$, respectively.
• Figure 5: The manipulators with different stiffness selections have varying reachable workspace.