Learning to Explore using Active Neural SLAM
This addresses the challenge of efficient exploration in robotics and AI for applications like navigation, but it is incremental as it builds on existing classical and learning-based methods.
The paper tackles the problem of exploring 3D environments by proposing Active Neural SLAM, a modular and hierarchical approach that combines classical path planners with learned SLAM and policies, resulting in effectiveness over past methods and winning the CVPR 2019 Habitat PointGoal Navigation Challenge.
This work presents a modular and hierarchical approach to learn policies for exploring 3D environments, called `Active Neural SLAM'. Our approach leverages the strengths of both classical and learning-based methods, by using analytical path planners with learned SLAM module, and global and local policies. The use of learning provides flexibility with respect to input modalities (in the SLAM module), leverages structural regularities of the world (in global policies), and provides robustness to errors in state estimation (in local policies). Such use of learning within each module retains its benefits, while at the same time, hierarchical decomposition and modular training allow us to sidestep the high sample complexities associated with training end-to-end policies. Our experiments in visually and physically realistic simulated 3D environments demonstrate the effectiveness of our approach over past learning and geometry-based approaches. The proposed model can also be easily transferred to the PointGoal task and was the winning entry of the CVPR 2019 Habitat PointGoal Navigation Challenge.