Self-Supervised Deep Pose Corrections for Robust Visual Odometry

Abstract

We present a self-supervised deep pose correction (DPC) network that applies pose corrections to a visual odometry estimator to improve its accuracy. Instead of regressing inter-frame pose changes directly, we build on prior work that uses data-driven learning to regress pose corrections that account for systematic errors due to violations of modelling assumptions. Our self-supervised formulation removes any requirement for six-degrees-of-freedom ground truth and, in contrast to expectations, often improves overall navigation accuracy compared to a supervised approach. Through extensive experiments, we show that our self-supervised DPC network can significantly enhance the performance of classical monocular and stereo odometry estimators and substantially out-performs state-of-the-art learning-only approaches.

arXiv commentIn Proceedings of the IEEE International Conference on Robotics and Automation (ICRA'20), Paris, France, May 31 - Jun. 4, 2020

References in corpus (11)

• Adam: A Method for Stochastic Optimization
• Past, Present, and Future of Simultaneous Localization And Mapping: Towards the Robust-Perception Age
• A guide to convolution arithmetic for deep learning
• DeepVO: Towards End-to-End Visual Odometry with Deep Recurrent Convolutional Neural Networks
• SfM-Net: Learning of Structure and Motion from Video
• A micro Lie theory for state estimation in robotics
• Backprop KF: Learning Discriminative Deterministic State Estimators
• Does computer vision matter for action?
• DPC-Net: Deep Pose Correction for Visual Localization
• Reducing Drift in Visual Odometry by Inferring Sun Direction Using a Bayesian Convolutional Neural Network
• Robustness Meets Deep Learning: An End-to-End Hybrid Pipeline for Unsupervised Learning of Egomotion

Cited by in corpus (6)

• Perception and Navigation in Autonomous Systems in the Era of Learning: A Survey
• Self-Supervised Scale Recovery for Monocular Depth and Egomotion Estimation
• Deep Online Correction for Monocular Visual Odometry
• Salient Sparse Visual Odometry With Pose-Only Supervision
• On the Coupling of Depth and Egomotion Networks for Self-Supervised Structure from Motion
• Robot Perception enables Complex Navigation Behavior via Self-Supervised Learning