Industrial robot manipulators are playing a significant role in modern manufacturing industries. Though peg-in-hole assembly is a common industrial task that has been extensively researched, safely solving complex, high-precision assembly in an unstructured environment remains an open problem. Reinforcement-learning (RL) methods have proven to be successful in autonomously solving manipulation tasks. However, RL is still not widely adopted in real robotic systems because working with real hardware entails additional challenges, especially when using position-controlled manipulators. The main contribution of this work is a learning-based method to solve peg-in-hole tasks with hole-position uncertainty. We propose the use of an off-policy, model-free reinforcement-learning method, and we bootstraped the training speed by using several transfer-learning techniques (sim2real) and domain randomization. Our proposed learning framework for position-controlled robots was extensively evaluated in contact-rich insertion tasks in a variety of environments.
This paper presents a method for planning motions of a flexible objects based on precise simulation using Finite Element Method (FEM). The proposed method is applied to ring-shape objects manipulated by robot arms, which is often seen in various applications. Since large deformation is implied, assembly planning with realistic simulation is important to ensure task efficiency for the robot and also to avoid damage of the object. We first verify that the behavior of a ring-shape object by dual-arm manipulation is well predicted using FEM model of bent beam through a simulation along the trajectory computed by optimization-based motion planning previously reported. Next, a precise FEM model is integrated into optimization to compute a trajectory of robot hands minimizing the deformation energy as well as satisfying such criteria as collision avoidance and smoothness. Since the direct computation leads huge computational cost, we present a realistic formula which transforms the planning problem into the static equilibrium problem of several FEM models located along the trajectory. Simulation results show that the proposed method is promising for such assembly tasks requiring large deformation.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.