PN-WSNA-Based Eye–Hand–Leg Coordination with a FIRA HuroCup Basketball Game

Biped humanoid robots present perception and locomotion capabilities similar to those of human beings. Eye-hand-leg coordination is a basic and required capability for a healthy person, and eye-hand-leg coordination performs various skill-oriented tasks. In this paper, the skill-oriented tasks for playing a basketball game were implemented for a child-size humanoid robot (HuroEvolution) through the use of eye-hand-leg coordination. The “eye” function is responsible for active perceptions in environments, the “hand” function is for gripping and manipulating objects to perform specific tasks, and the “leg” function is used to transport the biped robot to a desired location for the execution of tasks. Coordination among the eye, hand, and leg was achieved using predefined scenarios for various skill-oriented tasks. To utilize flexible and systematic skill-oriented task implementations, a Petri net-based wireless sensor node architecture (PN-WSNA) was employed in this project to define the task scenarios. The control scenario of eye-hand-leg coordination was realized by aggregating fundamental PN-WSNA scenario modules for fast realizations. The coordination model can be further evaluated theoretically to investigate its propensity for avoiding system deadlocks. Finally, the skill-oriented eye-hand-leg coordination tasks for realizing the Federation of International Robot-Soccer Association HuroCup League basketball games were successfully demonstrated to validate our approaches.