Visual Representations for Semantic Target Driven Navigation

What is a good visual representation for navigation? We study this question in the context of semantic visual navigation, which is the problem of a robot finding its way through a previously unseen environment to a target object, e. g. go to the refrigerator. Instead of acquiring a metric semantic map of an environment and using planning for navigation, our approach learns navigation policies on top of representations that capture spatial layout and semantic contextual cues. We propose to use semantic segmentation and detection masks as observations obtained by state-of-the-art computer vision algorithms and use a deep network to learn the navigation policy. The availability of equitable representations in simulated environments enables joint training using real and simulated data and alleviates the need for domain adaptation or domain randomization commonly used to tackle the sim-to-real transfer of the learned policies. Both the representation and the navigation policy can be readily applied to real non-synthetic environments as demonstrated on the Active Vision Dataset [1]. Our approach successfully gets to the target in 54% of the cases in unexplored environments, compared to 46% for a non-learning based approach, and 28% for a learning-based baseline.

Authors

• Arsalan Mousavian
• Alexander Toshev
• Marek Fiser
• Jana Kosecka
• Ayzaan Wahid
• James Davidson

Keywords

• Navigation
• Visualization
• Semantics
• Training
• Adaptation models
• Robots
• Task analysis
• Visual Representation
• Deep Network
• Simulated Data
• Simulation Environment
• Semantic Segmentation
• Domain Adaptation
• Contextual Cues
• Policy Learning
• Machine Vision
• Deep Learning
• Convolutional Layers
• Object Detection
• Shortest Path
• Raw Images
• Visual Observation
• Object Location
• Path Planning
• Depth Images
• Optimal Path
• Semantic Representations
• Imitation Learning
• Raw Depth
• Visual Question Answering
• Navigation Strategies
• Synthetic Environment
• Raw Observations
• Reinforcement Learning Scheme
• Environmental Knowledge
• Collision