Visual-Inertial Target Tracking and Motion Planning for UAV-based Radiation Detection
This addresses the challenge of radiation detection for security or monitoring applications using UAVs, but it appears incremental as it builds on existing motion planning and tracking methods.
The paper tackled the problem of detecting radioactive material in transit using a UAV with minimal sensing, by developing a motion planning framework that integrates visual-inertial localization and target tracking to generate safe, obstacle-avoiding paths that converge to a moving target, validated through realistic simulations in Gazebo.
This paper addresses the problem of detecting radioactive material in transit using an UAV of minimal sensing capability, where the objective is to classify the target's radioactivity as the vehicle plans its paths through the workspace while tracking the target for a short time interval. To this end, we propose a motion planning framework that integrates tightly-coupled visual-inertial localization and target tracking. In this framework,the 3D workspace is known, and this information together with the UAV dynamics, is used to construct a navigation function that generates dynamically feasible, safe paths which avoid obstacles and provably converge to the moving target. The efficacy of the proposed approach is validated through realistic simulations in Gazebo.