Differentiable SLAM-net: Learning Particle SLAM for Visual Navigation
This work addresses robust visual navigation for robots in challenging conditions, representing a strong incremental advance with specific performance gains.
The paper tackles the challenge of visual robot navigation in unseen indoor environments by introducing Differentiable SLAM-net, which encodes a particle filter SLAM algorithm in a differentiable graph and learns neural components end-to-end, resulting in a 37% to 64% success rate improvement on the Habitat Challenge 2020 PointNav task.
Simultaneous localization and mapping (SLAM) remains challenging for a number of downstream applications, such as visual robot navigation, because of rapid turns, featureless walls, and poor camera quality. We introduce the Differentiable SLAM Network (SLAM-net) along with a navigation architecture to enable planar robot navigation in previously unseen indoor environments. SLAM-net encodes a particle filter based SLAM algorithm in a differentiable computation graph, and learns task-oriented neural network components by backpropagating through the SLAM algorithm. Because it can optimize all model components jointly for the end-objective, SLAM-net learns to be robust in challenging conditions. We run experiments in the Habitat platform with different real-world RGB and RGB-D datasets. SLAM-net significantly outperforms the widely adapted ORB-SLAM in noisy conditions. Our navigation architecture with SLAM-net improves the state-of-the-art for the Habitat Challenge 2020 PointNav task by a large margin (37% to 64% success). Project website: http://sites.google.com/view/slamnet