Target Tracking with an Unmanned Aerial Vehicle Using Visual Servoing

Unmanned aerial vehicles (UAV), frequently referred to as drones, are rapidly being used for military and civilian applications. Surveillance and target tracking are two of the newer applications for unmanned systems. Drones are used to track targets, however the pilot often has to operate the drone while also observing the area. The drone must be made autonomous during tracking in order to prevent damage and free up the operator to concentrate on the situation. Using the AR. Drone 2.0, we created a prototype for target tracking via visual servoing. To accomplish this, we used the Robot Operating System (ROS) development platform and the Target Tracking Library (OpenTLD) to develop proportional, integral, and derivative (PID) controllers. This platform enabled us to model our solution in the gazebo simulator and then do tracking tests in a real environment. The tracking is carried out without the use of GPS, using only the drone's front camera, and the computational work is done on a laptop.