Zero-Shot Reinforcement Learning on Graphs for Autonomous Exploration Under Uncertainty

• URL: http://arxiv.org/abs/2105.04758v1
• Date: Tue, 11 May 2021 02:42:17 GMT
• Title: Zero-Shot Reinforcement Learning on Graphs for Autonomous Exploration Under Uncertainty
• Authors: Fanfei Chen, Paul Szenher, Yewei Huang, Jinkun Wang, Tixiao Shan, Shi Bai, Brendan Englot
• Abstract summary: We present a framework for self-learning a high-performance exploration policy in a single simulation environment. We propose a novel approach that uses graph neural networks in conjunction with deep reinforcement learning.
• Score: 6.42522897323111
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: This paper studies the problem of autonomous exploration under localization uncertainty for a mobile robot with 3D range sensing. We present a framework for self-learning a high-performance exploration policy in a single simulation environment, and transferring it to other environments, which may be physical or virtual. Recent work in transfer learning achieves encouraging performance by domain adaptation and domain randomization to expose an agent to scenarios that fill the inherent gaps in sim2sim and sim2real approaches. However, it is inefficient to train an agent in environments with randomized conditions to learn the important features of its current state. An agent can use domain knowledge provided by human experts to learn efficiently. We propose a novel approach that uses graph neural networks in conjunction with deep reinforcement learning, enabling decision-making over graphs containing relevant exploration information provided by human experts to predict a robot's optimal sensing action in belief space. The policy, which is trained only in a single simulation environment, offers a real-time, scalable, and transferable decision-making strategy, resulting in zero-shot transfer to other simulation environments and even real-world environments.

Related papers

• Sim-to-Real Transfer of Deep Reinforcement Learning Agents for Online Coverage Path Planning [15.792914346054502]
  We tackle the challenge of sim-to-real transfer of reinforcement learning (RL) agents for coverage path planning ( CPP) We bridge the sim-to-real gap through a semi-virtual environment, including a real robot and real-time aspects, while utilizing a simulated sensor and obstacles. We find that a high inference frequency allows first-order Markovian policies to transfer directly from simulation, while higher-order policies can be fine-tuned to further reduce the sim-to-real gap.
  arXiv  Detail & Related papers  (2024-06-07T13:24:19Z)
• Learning to navigate efficiently and precisely in real environments [14.52507964172957]
  Embodied AI literature focuses on end-to-end agents trained in simulators like Habitat or AI-Thor. In this work we explore end-to-end training of agents in simulation in settings which minimize the sim2real gap.
  arXiv  Detail & Related papers  (2024-01-25T17:50:05Z)
• Planning for Learning Object Properties [117.27898922118946]
  We formalize the problem of automatically training a neural network to recognize object properties as a symbolic planning problem. We use planning techniques to produce a strategy for automating the training dataset creation and the learning process. We provide an experimental evaluation in both a simulated and a real environment.
  arXiv  Detail & Related papers  (2023-01-15T09:37:55Z)
• Imitation Learning for Generalizable Self-driving Policy with Sim-to-real Transfer [0.0]
  We present three Imitation Learning and two sim-to-real methods capable of achieving this task. A detailed comparison is provided on these techniques to highlight their advantages and disadvantages.
  arXiv  Detail & Related papers  (2022-06-22T01:36:14Z)
• Towards Autonomous Grading In The Real World [4.651327752886103]
  We aim to tackle the problem of autonomous grading, where a dozer is required to flatten an uneven area. We design both a realistic physical simulation and a scaled real prototype environment mimicking the real dozer dynamics and sensory information.
  arXiv  Detail & Related papers  (2022-06-13T12:21:20Z)
• Stochastic Coherence Over Attention Trajectory For Continuous Learning In Video Streams [64.82800502603138]
  This paper proposes a novel neural-network-based approach to progressively and autonomously develop pixel-wise representations in a video stream. The proposed method is based on a human-like attention mechanism that allows the agent to learn by observing what is moving in the attended locations. Our experiments leverage 3D virtual environments and they show that the proposed agents can learn to distinguish objects just by observing the video stream.
  arXiv  Detail & Related papers  (2022-04-26T09:52:31Z)
• Towards Optimal Strategies for Training Self-Driving Perception Models in Simulation [98.51313127382937]
  We focus on the use of labels in the synthetic domain alone. Our approach introduces both a way to learn neural-invariant representations and a theoretically inspired view on how to sample the data from the simulator. We showcase our approach on the bird's-eye-view vehicle segmentation task with multi-sensor data.
  arXiv  Detail & Related papers  (2021-11-15T18:37:43Z)
• Point Cloud Based Reinforcement Learning for Sim-to-Real and Partial Observability in Visual Navigation [62.22058066456076]
  Reinforcement Learning (RL) represents powerful tools to solve complex robotic tasks. RL does not work directly in the real-world, which is known as the sim-to-real transfer problem. We propose a method that learns on an observation space constructed by point clouds and environment randomization.
  arXiv  Detail & Related papers  (2020-07-27T17:46:59Z)
• Autonomous Exploration Under Uncertainty via Deep Reinforcement Learning on Graphs [5.043563227694137]
  We consider an autonomous exploration problem in which a range-sensing mobile robot is tasked with accurately mapping the landmarks in an a priori unknown environment efficiently in real-time. We propose a novel approach that uses graph neural networks (GNNs) in conjunction with deep reinforcement learning (DRL), enabling decision-making over graphs containing exploration information to predict a robot's optimal sensing action in belief space.
  arXiv  Detail & Related papers  (2020-07-24T16:50:41Z)
• Guided Uncertainty-Aware Policy Optimization: Combining Learning and Model-Based Strategies for Sample-Efficient Policy Learning [75.56839075060819]
  Traditional robotic approaches rely on an accurate model of the environment, a detailed description of how to perform the task, and a robust perception system to keep track of the current state. reinforcement learning approaches can operate directly from raw sensory inputs with only a reward signal to describe the task, but are extremely sample-inefficient and brittle. In this work, we combine the strengths of model-based methods with the flexibility of learning-based methods to obtain a general method that is able to overcome inaccuracies in the robotics perception/actuation pipeline.
  arXiv  Detail & Related papers  (2020-05-21T19:47:05Z)
• The Chef's Hat Simulation Environment for Reinforcement-Learning-Based Agents [54.63186041942257]
  We propose a virtual simulation environment that implements the Chef's Hat card game, designed to be used in Human-Robot Interaction scenarios. This paper provides a controllable and reproducible scenario for reinforcement-learning algorithms.
  arXiv  Detail & Related papers  (2020-03-12T15:52:49Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.