PROBE: Predictive robust estimation for visual-inertial navigation

Navigation in unknown, chaotic environments continues to present a significant challenge for the robotics community. Lighting changes, self-similar textures, motion blur, and moving objects are all considerable stumbling blocks for state-of-the-art vision-based navigation algorithms. In this paper we present a novel technique for improving localization accuracy within a visual-inertial navigation system (VINS). We make use of training data to learn a model for the quality of visual features with respect to localization error in a given environment. This model maps each visual observation from a predefined prediction space of visual-inertial predictors onto a scalar weight, which is then used to scale the observation covariance matrix. In this way, our model can adjust the influence of each observation according to its quality. We discuss our choice of predictors and report substantial reductions in localization error on 4 km of data from the KITTI dataset, as well as on experimental datasets consisting of 700 m of indoor and outdoor driving on a small ground rover equipped with a Skybotix VI-Sensor.

Authors

• Valentin Peretroukhin
• Lee E. Clement
• Matthew Giamou
• Jonathan Kelly

Keywords

• Visualization
• Navigation
• Probes
• Cameras
• Robot sensing systems
• Robustness
• Uncertainty
• Visual-inertial Navigation
• Training Data
• Covariance Matrix
• Visual Features
• Quality Characteristics
• Navigation System
• Motion Blur
• Predictors Of Choice
• KITTI Dataset
• Predictor Space
• Scalar Weights
• Improve Localization Accuracy
• Root Mean Square Error
• Small Region
• Angular Velocity
• Point Cloud
• Linear Accelerator
• Optical Flow
• Feature Matching
• Pose Estimation
• Random Sample Consensus
• Left Camera
• Outlier Rejection
• Stereo Camera
• Root Mean Square Error Of Cross-validation
• Image Coordinates
• Metric Scale
• Feature Tracking
• Sensor Noise
• Reprojection Error