Autonomous Coverage Path Planning and Navigation Control System for Search Operations using a UAV
This paper introduces an autonomous control system to search for a designated target by the use of UAV. The system calculates a coverage path plan for a designated area around a provided location, then navigates through it until the specified target has been found or the entire area has been covered. We combined NED (North-East-Down) coordinates, the relative FRU (Front-Right-Up) position of the UAV and GPS location to plan and execute movements. Using the motion algorithm, we calculate and generate the minimum required way-points to completely cover the desired area. A image processing technique for object detection and a camera facing downwards continuously searches for the designated target. When the image detection system has found the target, the navigation is halted and the target location is reported. We conducted experiments in real environment to test our object detection model. A 3d simulation environment using Gazebo was also implemented for extensive testing in similar conditions to the real world and in multiple scenarios.
Paper
Full text
Autonomous Coverage Path Planning and Navigation Control System for Search Operations using a UAV
Semantic Scholar · Engineering · 2019
Abstract
This paper introduces an autonomous control system to search for a designated target by the use of UAV. The system calculates a coverage path plan for a designated area around a provided location, then navigates through it until the specified target has been found or the entire area has been covered. We combined NED (North-East-Down) coordinates, the relative FRU (Front-Right-Up) position of the UAV and GPS location to plan and execute movements. Using the motion algorithm, we calculate and generate the minimum required way-points to completely cover the desired area. A image processing technique for object detection and a camera facing downwards continuously searches for the designated target. When the image detection system has found the target, the navigation is halted and the target location is reported. We conducted experiments in real environment to test our object detection model. A 3d simulation environment using Gazebo was also implemented for extensive testing in similar conditions to the real world and in multiple scenarios.