A Reactive Method for Collision Avoidance in Industrial Environments

• Published in: Journal of intelligent & robotic systems Vol. 84; no. 1-4; pp. 745 - 758
• Main Authors: Alejo, D., Cobano, J. A., Heredia, G., Ollero, A.
• Format: Journal Article
• Language: English
• Published: Dordrecht Springer Netherlands 01.12.2016; Springer Nature B.V
• Subjects: Aerials; Algorithms; Artificial Intelligence; Collision avoidance; Computer simulation; Control; Electrical Engineering; Engineering; Fatigue tests; Mechanical Engineering; Mechatronics; Obstacles; Real time; Robotics; Vehicles; Real-time; Cooperative; Experimental; Collision; Safety; Reciprocal; Reactive; Multi-UAS; Avoidance; Quadrotor
• ISSN: 0921-0296; 1573-0409
• DOI: 10.1007/s10846-016-0359-7

This paper presents a reactive method for collision avoidance with multiple aerial vehicles that has been applied in real time considering industrial environments. The proposed method is based on the 3D-Optimal Reciprocal Collision Avoidance algorithm. The main contribution of the proposed method is that it takes into consideration 3D modeled static obstacles. Therefore, it has been successfully applied in realistic industrial environments with the presence of complex static obstacles. Considerations of dynamic constraints of the aerial vehicles have been added. The algorithm has been integrated in ROS framework and tested in simulation. Several simulations with up to eight aerial vehicles have been performed, including long endurance cooperative missions. Finally, the second main contribution consists of the evaluation of several real experiments with up to four aerial vehicles which have been carried out in the testbed of the Center for Advanced Technologies (CATEC) facilities. The aerial vehicles flew in the presence of static obstacles and avoided potential collisions by modifying the planned trajectories in real-time.