Matches in SemOpenAlex for { <https://semopenalex.org/work/W4300938316> ?p ?o ?g. }
Showing items 1 to 69 of
69
with 100 items per page.
- W4300938316 abstract "Deep reinforcement learning has emerged as a promising and powerful technique for automatically acquiring control policies that can process raw sensory inputs, such as images, and perform complex behaviors. However, extending deep RL to real-world robotic tasks has proven challenging, particularly in safety-critical domains such as autonomous flight, where a trial-and-error learning process is often impractical. In this paper, we explore the following question: can we train vision-based navigation policies entirely in simulation, and then transfer them into the real world to achieve real-world flight without a single real training image? We propose a learning method that we call CAD$^2$RL, which can be used to perform collision-free indoor flight in the real world while being trained entirely on 3D CAD models. Our method uses single RGB images from a monocular camera, without needing to explicitly reconstruct the 3D geometry of the environment or perform explicit motion planning. Our learned collision avoidance policy is represented by a deep convolutional neural network that directly processes raw monocular images and outputs velocity commands. This policy is trained entirely on simulated images, with a Monte Carlo policy evaluation algorithm that directly optimizes the network's ability to produce collision-free flight. By highly randomizing the rendering settings for our simulated training set, we show that we can train a policy that generalizes to the real world, without requiring the simulator to be particularly realistic or high-fidelity. We evaluate our method by flying a real quadrotor through indoor environments, and further evaluate the design choices in our simulator through a series of ablation studies on depth prediction. For supplementary video see: https://youtu.be/nXBWmzFrj5s" @default.
- W4300938316 created "2022-10-04" @default.
- W4300938316 creator A5006431582 @default.
- W4300938316 creator A5026322200 @default.
- W4300938316 date "2016-11-13" @default.
- W4300938316 modified "2023-09-25" @default.
- W4300938316 title "CAD2RL: Real Single-Image Flight without a Single Real Image" @default.
- W4300938316 doi "https://doi.org/10.48550/arxiv.1611.04201" @default.
- W4300938316 hasPublicationYear "2016" @default.
- W4300938316 type Work @default.
- W4300938316 citedByCount "1" @default.
- W4300938316 countsByYear W43009383162023 @default.
- W4300938316 crossrefType "posted-content" @default.
- W4300938316 hasAuthorship W4300938316A5006431582 @default.
- W4300938316 hasAuthorship W4300938316A5026322200 @default.
- W4300938316 hasBestOaLocation W43009383161 @default.
- W4300938316 hasConcept C108583219 @default.
- W4300938316 hasConcept C111919701 @default.
- W4300938316 hasConcept C121704057 @default.
- W4300938316 hasConcept C154945302 @default.
- W4300938316 hasConcept C205711294 @default.
- W4300938316 hasConcept C2776459999 @default.
- W4300938316 hasConcept C2780864053 @default.
- W4300938316 hasConcept C31972630 @default.
- W4300938316 hasConcept C38652104 @default.
- W4300938316 hasConcept C41008148 @default.
- W4300938316 hasConcept C44154836 @default.
- W4300938316 hasConcept C65909025 @default.
- W4300938316 hasConcept C76155785 @default.
- W4300938316 hasConcept C79403827 @default.
- W4300938316 hasConcept C81363708 @default.
- W4300938316 hasConcept C82990744 @default.
- W4300938316 hasConcept C97541855 @default.
- W4300938316 hasConcept C98045186 @default.
- W4300938316 hasConceptScore W4300938316C108583219 @default.
- W4300938316 hasConceptScore W4300938316C111919701 @default.
- W4300938316 hasConceptScore W4300938316C121704057 @default.
- W4300938316 hasConceptScore W4300938316C154945302 @default.
- W4300938316 hasConceptScore W4300938316C205711294 @default.
- W4300938316 hasConceptScore W4300938316C2776459999 @default.
- W4300938316 hasConceptScore W4300938316C2780864053 @default.
- W4300938316 hasConceptScore W4300938316C31972630 @default.
- W4300938316 hasConceptScore W4300938316C38652104 @default.
- W4300938316 hasConceptScore W4300938316C41008148 @default.
- W4300938316 hasConceptScore W4300938316C44154836 @default.
- W4300938316 hasConceptScore W4300938316C65909025 @default.
- W4300938316 hasConceptScore W4300938316C76155785 @default.
- W4300938316 hasConceptScore W4300938316C79403827 @default.
- W4300938316 hasConceptScore W4300938316C81363708 @default.
- W4300938316 hasConceptScore W4300938316C82990744 @default.
- W4300938316 hasConceptScore W4300938316C97541855 @default.
- W4300938316 hasConceptScore W4300938316C98045186 @default.
- W4300938316 hasLocation W43009383161 @default.
- W4300938316 hasLocation W43009383162 @default.
- W4300938316 hasOpenAccess W4300938316 @default.
- W4300938316 hasPrimaryLocation W43009383161 @default.
- W4300938316 hasRelatedWork W1803059841 @default.
- W4300938316 hasRelatedWork W2557924869 @default.
- W4300938316 hasRelatedWork W2794771666 @default.
- W4300938316 hasRelatedWork W2891001608 @default.
- W4300938316 hasRelatedWork W2980953096 @default.
- W4300938316 hasRelatedWork W3081997306 @default.
- W4300938316 hasRelatedWork W3102636071 @default.
- W4300938316 hasRelatedWork W4283017422 @default.
- W4300938316 hasRelatedWork W4298315644 @default.
- W4300938316 hasRelatedWork W4321512589 @default.
- W4300938316 isParatext "false" @default.
- W4300938316 isRetracted "false" @default.
- W4300938316 workType "article" @default.