Benchmarking Classic and Learned Navigation in Complex 3D Environments
arXiv:1901.10915
Abstract
Navigation research is attracting renewed interest with the advent of learning-based methods. However, this new line of work is largely disconnected from well-established classic navigation approaches. In this paper, we take a step towards coordinating these two directions of research. We set up classic and learning-based navigation systems in common simulated environments and thoroughly evaluate them in indoor spaces of varying complexity, with access to different sensory modalities. Additionally, we measure human performance in the same environments. We find that a classic pipeline, when properly tuned, can perform very well in complex cluttered environments. On the other hand, learned systems can operate more robustly with a limited sensor suite. Overall, both approaches are still far from human-level performance.
Added CNN-Monodepth and OpenCV Stereo agents
References in corpus (10)
- Empirical Evaluation of Rectified Activations in Convolutional Network
- On Evaluation of Embodied Navigation Agents
- Learning to Navigate in Complex Environments
- Matterport3D: Learning from RGB-D Data in Indoor Environments
- Reinforcement Learning with Unsupervised Auxiliary Tasks
- Building Generalizable Agents with a Realistic and Rich 3D Environment
- Control of Memory, Active Perception, and Action in Minecraft
- MINOS: Multimodal Indoor Simulator for Navigation in Complex Environments
- Learning to Act by Predicting the Future
- Playing Doom with SLAM-Augmented Deep Reinforcement Learning
Cited by in corpus (17)
- DD-PPO: Learning Near-Perfect PointGoal Navigators from 2.5 Billion Frames
- Learning hierarchical relationships for object-goal navigation
- Learning to Set Waypoints for Audio-Visual Navigation
- A Survey of Embodied AI: From Simulators to Research Tasks
- Integrating Egocentric Localization for More Realistic Point-Goal Navigation Agents
- Sparse Graphical Memory for Robust Planning
- To Learn or Not to Learn: Analyzing the Role of Learning for Navigation in Virtual Environments
- VideoNavQA: Bridging the Gap between Visual and Embodied Question Answering
- Bayesian Relational Memory for Semantic Visual Navigation
- SoundSpaces: Audio-Visual Navigation in 3D Environments
- Learning Object-conditioned Exploration using Distributed Soft Actor Critic
- SGoLAM: Simultaneous Goal Localization and Mapping for Multi-Object Goal Navigation
- Learning to Map for Active Semantic Goal Navigation
- MaAST: Map Attention with Semantic Transformersfor Efficient Visual Navigation
- Efficient Robotic Object Search via HIEM: Hierarchical Policy Learning with Intrinsic-Extrinsic Modeling
- Semantic Audio-Visual Navigation
- Point Cloud Based Reinforcement Learning for Sim-to-Real and Partial Observability in Visual Navigation