A Low-Cost Relative Positioning Method for UAV/UGV Coordinated Heterogeneous System Based on Visual-Lidar Fusion

Abstract

Unmanned Ground Vehicles (UGVs) and Unmanned Aerial Vehicles (UAVs) are commonly used for various purposes, and their cooperative systems have been developed to enhance their capabilities. However, tracking and interacting with dynamic UAVs poses several challenges, including limitations of traditional radar and visual systems, and the need for the real-time monitoring of UAV positions. To address these challenges, a low-cost method that uses LiDAR (Light Detection and Ranging) and RGB-D cameras to detect and track UAVs in real time has been proposed. This method relies on a learning model and a linear Kalman filter, and has demonstrated satisfactory estimation accuracy using only CPU (Central Processing Unit)- in GPS (Global Positioning System)-denied environments without any prior information.

Keywords

• deep learning
• UAV tracking
• object detection
• linear Kalman filter
• LiDAR-inertial odometry