Array (Mar 2023)
Kinematic modes of vision-based heterogeneous UAV-AGV system
Abstract
UAV-AGV (Unmanned Aerial Vehicle-Autonomous Ground Vehicle) heterogeneous multi-agent system is gaining attention of researchers and became the topic of interest. UAV-AGV system is more advantageous than individual robots as both mobile platforms benefit from independent capabilities. Vision-based technique is the most researched method to build collaborative connection between both agents as it can function in GPS inaccessible areas. Take-off, tracking, and landing are basic tasks of the system. Vision-based solutions for these tasks are given most attention by various researchers so far. However, complex collaborative kinematics between agents are necessary to accomplish the complicated tasks which requires further investigation of the system. Moreover, a detailed kinematic model is necessary to better describe a robotic system which still needs to be developed to the vision-based UAV-AGV system. Kinematic analysis of the system is performed in the present work to address these issues. A differential form of mathematical representation is obtained for the vision-based collaboration between UAV and AGV. Forward and inverse kinematic notions are introduced for the system. Three kinematic modes of the system are identified and explained based on the kinematic model. Third mode reveals an interesting relative motion and novel behaviour of the system. Further step is taken to implement sliding mode-based control theory to develop novel kinematic controllers for the motion of UAV using each kinematic mode. Kinematic controllers are validated with thorough simulations. The performance of the controllers is acceptable and may be applied to activate the different modes of the system depending on the mission requirement.
