An intuitive tele-collaboration interface exploring laser-based interaction and behavior trees
Davide Torielli, Luca Muratore, Nikos G. Tsagarakis · Robotics and Autonomous Systems · 2025
The recent advancements in the development of robotic systems that offer advanced loco-manipulation capabilities have opened new opportunities in the employment of such platforms in various domains. However, despite the increased range of offered capabilities, the collaboration with these robotic platforms to execute tasks through common human–robot interaction interfaces is still an open challenge. In this article, we present a novel human–robot interaction interface that permits to intuitively command and control the manipulation and the locomotion abilities of the robot by exploring a visual servoing guidance method realized with a laser emitter device. By pointing the laser to locations and objects in the environment where the robot is operating, the operator is able to command even highly articulated robots intuitively and efficiently. The detection of the laser projection is performed by a neural network that provides robust and real-time tracking of laser spot. Combined with the responsiveness of the laser detection, a Behavior Trees-based motion planner is employed to reactively select and generate the autonomous robot motions to reach the indicated target. This combination allows the operator to communicate goal locations and paths to follow without requiring prior knowledge of the system, and without worrying about the generation of the potential complex loco-manipulation robot actions. The effectiveness of the proposed interface is demonstrated with the CENTAURO robot, a hybrid leg-wheel platform with an anthropomorphic upper body, exploiting its abilities to accomplish a number of locomotion and manipulation tasks.