XIRVIO: Critic-guided Iterative Refinement for Visual-Inertial Odometry with Explainable Adaptive Weighting

Abstract-- We introduce XIRVIO, a transformer-based Generative Adversarial Network (GAN) framework for monocular visual inertial odometry (VIO). By taking sequences of imag es and 6-DoF inertial measurements as inputs, XIRVIO's generator predicts pose trajectories through an iterative refine ment process which are then evaluated by the critic to select the iteration with the optimised prediction. Additionally, th e self-emergent adaptive sensor weighting reveals how XIRVIO attends to each sensory input based on contextual cues in the da ta, making it a promising approach for achieving explainabilit y in safety-critical VIO applications. Evaluations on the KI TTI dataset demonstrate that XIRVIO matches well-known state-of-the-art learning-based methods in terms of both transla tion and rotation errors. Accurate and reliable state estimation is fundamental to th e autonomy of robotic systems, but can be challenging when navigating cluttered indoor spaces, dynamic urban environ - ments, and unstructured natural terrains like forests [1]- [4]. VIO leverages the complementary strengths of cameras and inertial measurement units (IMUs) to estimate the camera motion, but its performance is inherently tied to the reliab ility of each sensor under varying conditions.