Unifying Map and Landmark Based Representations for Visual Navigation
arXiv:1712.08125
Abstract
This works presents a formulation for visual navigation that unifies map based spatial reasoning and path planning, with landmark based robust plan execution in noisy environments. Our proposed formulation is learned from data and is thus able to leverage statistical regularities of the world. This allows it to efficiently navigate in novel environments given only a sparse set of registered images as input for building representations for space. Our formulation is based on three key ideas: a learned path planner that outputs path plans to reach the goal, a feature synthesis engine that predicts features for locations along the planned path, and a learned goal-driven closed loop controller that can follow plans given these synthesized features. We test our approach for goal-driven navigation in simulated real world environments and report performance gains over competitive baseline approaches.
Project page with videos: https://s-gupta.github.io/cmpl/
References in corpus (8)
- Adam: A Method for Stochastic Optimization
- Learning to Navigate in Complex Environments
- Unsupervised Learning of Depth and Ego-Motion from Video
- Target-driven Visual Navigation in Indoor Scenes using Deep Reinforcement Learning
- Neural SLAM: Learning to Explore with External Memory
- Neural Map: Structured Memory for Deep Reinforcement Learning
- Memory Augmented Control Networks
- Learning Transferable Policies for Monocular Reactive MAV Control
Cited by in corpus (34)
- On Evaluation of Embodied Navigation Agents
- Neural SLAM: Learning to Explore with External Memory
- A Survey of Deep Network Solutions for Learning Control in Robotics: From Reinforcement to Imitation
- Learning to Navigate in Cities Without a Map
- Embodied Visual Navigation with Automatic Curriculum Learning in Real Environments
- Emergence of Exploratory Look-Around Behaviors through Active Observation Completion
- Learning Deployable Navigation Policies at Kilometer Scale from a Single Traversal
- Learning to Set Waypoints for Audio-Visual Navigation
- A Survey of Embodied AI: From Simulators to Research Tasks
- Causal Navigation by Continuous-time Neural Networks
- Path Planning using Neural A* Search
- Guided Feature Transformation (GFT): A Neural Language Grounding Module for Embodied Agents
- A Behavioral Approach to Visual Navigation with Graph Localization Networks
- Scene Memory Transformer for Embodied Agents in Long-Horizon Tasks
- Learning models for visual 3D localization with implicit mapping
- From Seeing to Moving: A Survey on Learning for Visual Indoor Navigation (VIN)
- To Learn or Not to Learn: Analyzing the Role of Learning for Navigation in Virtual Environments
- Learning to Navigate in Indoor Environments: from Memorizing to Reasoning
- Topological Planning with Transformers for Vision-and-Language Navigation
- Learning to Visually Navigate in Photorealistic Environments Without any Supervision
- MVP: Unified Motion and Visual Self-Supervised Learning for Large-Scale Robotic Navigation
- Graph Attention Memory for Visual Navigation
- Bayesian Relational Memory for Semantic Visual Navigation
- SoundSpaces: Audio-Visual Navigation in 3D Environments
- Scaling Local Control to Large-Scale Topological Navigation
- Gated Path Planning Networks
- SASRA: Semantically-aware Spatio-temporal Reasoning Agent for Vision-and-Language Navigation in Continuous Environments
- Cross-View Policy Learning for Street Navigation
- MaAST: Map Attention with Semantic Transformersfor Efficient Visual Navigation
- Deep Visual MPC-Policy Learning for Navigation
- Learning Autonomous Exploration and Mapping with Semantic Vision
- Egocentric Spatial Memory
- Polyline Generative Navigable Space Segmentation for Autonomous Visual Navigation
- Sidekick Policy Learning for Active Visual Exploration