Abstract:Map based visual inertial localization is a crucial step to reduce the drift in state estimation of mobile robots. The underlying problem for localization is to estimate the pose from a set of 3D-2D feature correspondences, of which the main challenge is the presence of outliers, especially in changing environment. In this paper, we propose a robust solution based on efficient global optimization of the consensus maximization problem, which is insensitive to high percentage of outliers. We first introduce translation invariant measurements (TIMs) for both points and lines to decouple the consensus maximization problem into rotation and translation subproblems, allowing for a two-stage solver with reduced search space. Then we show that (i) the rotation can be estimated by minimizing TIMs using only 1-dimensional branch-and-bound (BnB), (ii) the translation can be estimated by running 1-dimensional search for each of the three axes with prioritized progressive voting. Compared with the popular randomized solver, our solver achieves deterministic global convergence without requiring an initial value. Furthermore, ours is exponentially faster compared with existing BnB based methods. Finally, our experiments on both simulation and real-world datasets demonstrate that the proposed method gives accurate pose estimation even in the presence of 90% outliers (only 2 inliers).

Pose, Velocity and Landmark Position Estimation Using IMU and Bearing Measurements

An almost globally convergent observer for visual SLAM without persistent excitation.

Globally Optimal Consensus Maximization for Robust Visual Inertial Localization in Point and Line Map.

Pose Estimation for Spherical Robot with Kinematics-Based Encoder-Inertial Odometry

Universal Global State Estimation for Inertial Navigation Systems

A Nonlinear Navigation Observer Using IMU and Generic Position Information

Inertial-aided State and Slope Estimation Using a Monocular Camera

Whole-body Pose Estimation in Physical Rider-Bicycle Interactions with a Monocular Camera and a Set of Wearable Gyroscopes

Tightly coupled Visual Inertial Odometry based on Artificial Landmarks

State Estimation Using Single Body-Frame Bearing Measurements

Body-Segment Orientation Estimation in Rider-Bicycle Interactions with an Un-Calibrated Monocular Camera and Wearable Gyroscopes

Angular and Linear Velocity Observer Design Using Vector and Landmark Measurements

Globally Optimal Consensus Maximization for Relative Pose Estimation With Known Gravity Direction

Invariant Cubature Kalman Filtering-Based Visual-Inertial Odometry for Robot Pose Estimation

A Monocular Odometer for a Quadrotor Using a Homography Model and Inertial Cues

Velocity-aided IMU-based Attitude Estimation

Pose estimation and mapping based on IMU and LiDAR

A Predictor-based Attitude and Position Estimation for Rigid Bodies Moving in Planar Space by using Delayed Landmark Measurements

LiDAR and Inertial Fusion for Pose Estimation by Non-linear Optimization

Angular velocity and linear acceleration measurement bias estimators for the rigid body system with global exponential convergence

On the Design of Hybrid Pose and Velocity-bias Observers on Lie Group SE(3)