Hierarchical Attention and Graph Neural Networks: Toward Drift-Free Pose Estimation
This addresses drift in pose estimation for robotics or autonomous systems, representing a novel method rather than an incremental improvement.
The paper tackles the problem of drift in 3D pose estimation by replacing traditional geometric registration and pose graph optimization with a learned model using hierarchical attention and graph neural networks, resulting in significant accuracy improvements, especially in rotational components, as tested on the KITTI Odometry dataset.
The most commonly used method for addressing 3D geometric registration is the iterative closet-point algorithm, this approach is incremental and prone to drift over multiple consecutive frames. The Common strategy to address the drift is the pose graph optimization subsequent to frame-to-frame registration, incorporating a loop closure process that identifies previously visited places. In this paper, we explore a framework that replaces traditional geometric registration and pose graph optimization with a learned model utilizing hierarchical attention mechanisms and graph neural networks. We propose a strategy to condense the data flow, preserving essential information required for the precise estimation of rigid poses. Our results, derived from tests on the KITTI Odometry dataset, demonstrate a significant improvement in pose estimation accuracy. This improvement is especially notable in determining rotational components when compared with results obtained through conventional multi-way registration via pose graph optimization. The code will be made available upon completion of the review process.