The field of localization and tracking is experiencing significant advancements, driven by the development of innovative methods and technologies. A key direction in this field is the integration of multiple sensors and modalities to improve accuracy and robustness. For example, the combination of inertial measurement units (IMUs), visual-inertial odometry (VIO), and Ultra Wideband (UWB) is enabling more accurate and consistent localization. Additionally, the use of machine learning and deep learning techniques is becoming increasingly popular, allowing for more efficient and adaptive processing of sensor data. Noteworthy papers in this area include ReNiL, which presents a Bayesian deep-learning framework for pedestrian localization, and CVIRO, which proposes a consistent and tightly-coupled visual-inertial-ranging odometry system. Other notable works include L2Calib, which introduces a reinforcement learning-based extrinsic calibration framework, and WPTrack, which proposes a Wi-Fi and pressure insole fusion system for single target tracking. These advancements have significant implications for various applications, including robotics, IoT, and smart homes.
Advances in Localization and Tracking
Sources
ReNiL: Relative Neural Inertial Locator with Any-Scale Bayesian Inference
Graph-based Robot Localization Using a Graph Neural Network with a Floor Camera and a Feature Rich Industrial Floor
L2Calib: $SE(3)$-Manifold Reinforcement Learning for Robust Extrinsic Calibration with Degenerate Motion Resilience
Differentiable Adaptive Kalman Filtering via Optimal Transport
Pinching-Antenna Systems (PASS)-based Indoor Positioning
WPTrack: A Wi-Fi and Pressure Insole Fusion System for Single Target Tracking
WiFi-based Global Localization in Large-Scale Environments Leveraging Structural Priors from osmAG
CVIRO: A Consistent and Tightly-Coupled Visual-Inertial-Ranging Odometry on Lie Groups