Object-Oriented Option Framework for Robotics Manipulation in Clutter

D. Magazzeni E. Karpas

3 Mei 2020

Modern robots are increasingly capable of performing “basic” activities such as localization, navigation, and motion planning. However, for a robot to be considered intelligent, we would like it to be able to automatically combine these capabilities in order to achieve a high-level goal. The field of automated planning (sometimes called AI planning) deals with automatically synthesizing plans that combine basic actions to achieve a high-level goal. In this article, we focus on the intersection of automated planning and robotics and discuss some of the challenges and tools available to employ automated planning in controlling robots. We review different types of planning formalisms and discuss their advantages and limitations, especially in the context of planning robot actions. We conclude with a brief guide aimed at helping roboticists choose the right planning model to endow a robot with planning capabilities.

Daniel Gebauer Lisa Heuss Gunther Reinhart

12 Oktober 2023

High product diversity, dynamic market conditions, and a lack of skilled workers are current challenges in manufacturing. Industrial robots autonomously planning and completing upcoming production tasks can help companies address these challenges. In this publication, we focus on autonomous task planning within industrial robotics and investigate how to facilitate the use of automated planning techniques from the field of artificial intelligence for this purpose. First, we present a novel methodology to automatically adapt abstractly modeled planning domains to the characteristics of individual application cases a user intends to implement. A planning domain is a formalized representation of the robot’s working environment that builds the basis for automated planning. Second, we integrate this approach into the procedure for developing skills-based industrial robotic applications to enable them to perform autonomous task planning. Finally, we demonstrate the use of the methodology within the application field kitting in two reference scenarios with a mobile robot and a stationary robot cell. Using our methodology, persons without expertise in automated planning can enable a robot for autonomous task planning without much extra effort.

Abhi Gupta P. Allen Feng Xu + 2 lainnya

10 Maret 2020

Recent advances in deep reinforcement learning (RL) have demonstrated its potential to learn complex robotic manipulation tasks. However, RL still requires the robot to collect a large amount of real-world experience. To address this problem, recent works have proposed learning from expert demonstrations (LfD), particularly via inverse reinforcement learning (IRL), given its ability to achieve robust performance with only a small number of expert demonstrations. Nevertheless, deploying IRL on real robots is still challenging due to the large number of robot experiences it requires. This paper aims to address this scalability challenge with a robust, sample-efficient, and general meta-IRL algorithm, SQUIRL, that performs a new but related long-horizon task robustly given only a single video demonstration. First, this algorithm bootstraps the learning of a task encoder and a task-conditioned policy using behavioral cloning (BC). It then collects real-robot experiences and bypasses reward learning by directly recovering a Q-function from the combined robot and expert trajectories. Next, this algorithm uses the learned Q-function to re-evaluate all cumulative experiences collected by the robot to improve the policy quickly. In the end, the policy performs more robustly (90%+ success) than BC on new tasks while requiring no experiences at test time. Finally, our real-robot and simulated experiments demonstrate our algorithm’s generality across different state spaces, action spaces, and vision-based manipulation tasks, e.g., pick-pour-place and pick-carry-drop.

Fidel Esquivel Estay Tizian Jermann H. Kolvenbach + 2 lainnya

20 September 2024

We introduce a novel strategy for multi-robot sorting of waste objects using Reinforcement Learning. Our focus lies on finding optimal picking strategies that facilitate an effective coordination of a multi-robot system, subject to maximizing the waste removal potential. We realize this by formulating the sorting problem as an OpenAI gym environment and training a neural network with a deep reinforcement learning algorithm. The objective function is set up to optimize the picking rate of the robotic system. In simulation, we draw a performance comparison to an intuitive combinatorial game theory-based approach. We show that the trained policies outperform the latter and achieve up to 16% higher picking rates. Finally, the respective algorithms are validated on a hardware setup consisting of a two-robot sorting station able to process incoming waste objects through pick-and-place operations.

Pooria Poorsarvi Tehrani Gavriel State Ritvik Singh + 13 lainnya

10 Januari 2023

We present Orbit, a unified and modular framework for robot learning powered by Nvidia Isaac Sim. It offers a modular design to easily and efficiently create robotic environments with photo-realistic scenes and high-fidelity rigid and deformable body simulation. With Orbit, we provide a suite of benchmark tasks of varying difficulty– from single-stage cabinet opening and cloth folding to multi-stage tasks such as room reorganization. To support working with diverse observations and action spaces, we include fixed-arm and mobile manipulators with different physically-based sensors and motion generators. Orbit allows training reinforcement learning policies and collecting large demonstration datasets from hand-crafted or expert solutions in a matter of minutes by leveraging GPU-based parallelization. In summary, we offer an open-sourced framework that readily comes with 16 robotic platforms, 4 sensor modalities, 10 motion generators, more than 20 benchmark tasks, and wrappers to 4 learning libraries. With this framework, we aim to support various research areas, including representation learning, reinforcement learning, imitation learning, and task and motion planning. We hope it helps establish interdisciplinary collaborations in these communities, and its modularity makes it easily extensible for more tasks and applications in the future.