IKEA Furniture Assembly Environment for Long-Horizon Complex Manipulation Tasks

Lee, Youngwoon; Hu, Edward S.; Lim, Joseph J.

doi:10.1109/icra48506.2021.9560986

Cited by 64 publications

(38 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Some of these tasks have been shown to be easily solvable by random search algorithms [41] and thus should not be considered as sufficiently difficult for comparing algorithms. Leveraging physics simulators, many environments have been proposed that involve robots, fixed in place, for object manipulation tasks of varying complexity [51,54,16,47,24,20,76,75,35]. Learning policies for robot manipulation is challenging, compounded by the exploration difficulty of the task and the sample inefficiency of RL algorithms.…”

Section: Related Workmentioning

confidence: 99%

CORA: Benchmarks, Baselines, and Metrics as a Platform for Continual Reinforcement Learning Agents

Powers¹,

Xing²,

Kolve³

et al. 2021

Preprint

View full text Add to dashboard Cite

Progress in continual reinforcement learning has been limited due to several barriers to entry: missing code, high compute requirements, and a lack of suitable benchmarks. In this work, we present CORA, a platform for Continual Reinforcement Learning Agents that provides benchmarks, baselines, and metrics in a single code package. The benchmarks we provide are designed to evaluate different aspects of the continual RL challenge, such as catastrophic forgetting, plasticity, ability to generalize, and sample-efficient learning. Three of the benchmarks utilize video game environments (Atari, Procgen, NetHack). The fourth benchmark, CHORES, consists of four different task sequences in a visually realistic home simulator, drawn from a diverse set of task and scene parameters. To compare continual RL methods on these benchmarks, we prepare three metrics in CORA: continual evaluation, forgetting, and zero-shot forward transfer. Finally, CORA includes a set of performant, open-source baselines of existing algorithms for researchers to use and expand on. We release CORA and hope that the continual RL community can benefit from our contributions, to accelerate the development of new continual RL algorithms.

show abstract

Section: Related Workmentioning

confidence: 99%

CORA: Benchmarks, Baselines, and Metrics as a Platform for Continual Reinforcement Learning Agents

Powers¹,

Xing²,

Kolve³

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…• Sawyer Assembly: The Sawyer arm needs to assemble the fourth leg (already attached to the gripper) of a table into the vacant hole while avoiding collisions to other three legs. This environment is built upon the IKEA furniture assembly environment [32].…”

Section: Environmentsmentioning

confidence: 99%

Distilling Motion Planner Augmented Policies into Visual Control Policies for Robot Manipulation

Liu¹,

Uppal²,

Sukhatme³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Learning complex manipulation tasks in realistic, obstructed environments is a challenging problem due to hard exploration in the presence of obstacles and high-dimensional visual observations. Prior work tackles the exploration problem by integrating motion planning and reinforcement learning. However, the motion planner augmented policy requires access to state information, which is often not available in the real-world settings. To this end, we propose to distill a state-based motion planner augmented policy to a visual control policy via (1) visual behavioral cloning to remove the motion planner dependency along with its jittery motion, and (2) vision-based reinforcement learning with the guidance of the smoothed trajectories from the behavioral cloning agent. We evaluate our method on three manipulation tasks in obstructed environments and compare it against various reinforcement learning and imitation learning baselines. The results demonstrate that our framework is highly sample-efficient and outperforms the state-of-the-art algorithms. Moreover, coupled with domain randomization, our policy is capable of zero-shot transfer to unseen environment settings with distractors. Code and videos are available at https://clvrai.com/mopa-pd.

show abstract

“…There have been many attempts that learn shape-to-shape matching algorithms in application-specific domains: furniture assembly [16,28,30], object assembly [1,31], and object packing [47]. Most of these assembly algorithms operate under the assumption that each shape corresponds to a recognizable semantic object part [16,28,30].…”

Section: Introductionmentioning

confidence: 99%

Neural Shape Mating: Self-Supervised Object Assembly with Adversarial Shape Priors

Chen¹,

Li²,

Turpin³

et al. 2022

Preprint

View full text Add to dashboard Cite

Neural Shape MatingNeural Shape Mating Solid Shape Mating Form Fitting Shape PairsPoint Clouds Mating Results Neural Shape Mating Shell Shape MatingFigure 1. Pairwise 3D geometric shape mating. Neural Shape Mating (NSM) takes as input the point clouds of a pair of shapes and predicts the mating configuration as output. NSM learns to mate shapes together with self-supervision and does not require semantic information or target shapes as guidance at test time. Our method can be applied to various shape mating settings, including solid shape mating (top row), shell shape mating (middle row), and form fitting (bottom row).

show abstract

IKEA Furniture Assembly Environment for Long-Horizon Complex Manipulation Tasks

Cited by 64 publications

References 31 publications

CORA: Benchmarks, Baselines, and Metrics as a Platform for Continual Reinforcement Learning Agents

CORA: Benchmarks, Baselines, and Metrics as a Platform for Continual Reinforcement Learning Agents

Distilling Motion Planner Augmented Policies into Visual Control Policies for Robot Manipulation

Neural Shape Mating: Self-Supervised Object Assembly with Adversarial Shape Priors

Contact Info

Product

Resources

About