This article presents a control strategy for guiding a herd toward a goal region using a multi-robot system. In herding tasks, sheep act as noncooperative agents, combining aggregation tendencies with stochastic movements that complicate the guidance process. To overcome this challenge, we propose a cooperative control approach in which robots position themselves around the herd to exert pressure and steer it toward the target. The robots coordinate to encircle the herd, prioritizing areas with higher risk of escape, while the center of the robot formation is guided toward the goal. The effectiveness of the proposed strategy is validated through simulations, comparisons with state-of-the-art methods and real robotic experiments. Beyond demonstrating improved containment and more efficient guidance, the results highlight the method’s ability to maintain order and cohesion within the group, adapt to dynamic herd configurations, respect inter-individual spacing, and achieve guidance in less time. These advantages position our strategy as a highly effective and scalable solution for advanced robotic herding tasks.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Adaptive Multi-robot Herding via Dynamic Risk-Aware Angular Repositioning

  • Leyre Remartinez,
  • Alejandro Perez-Yus,
  • Rosario Aragues

摘要

This article presents a control strategy for guiding a herd toward a goal region using a multi-robot system. In herding tasks, sheep act as noncooperative agents, combining aggregation tendencies with stochastic movements that complicate the guidance process. To overcome this challenge, we propose a cooperative control approach in which robots position themselves around the herd to exert pressure and steer it toward the target. The robots coordinate to encircle the herd, prioritizing areas with higher risk of escape, while the center of the robot formation is guided toward the goal. The effectiveness of the proposed strategy is validated through simulations, comparisons with state-of-the-art methods and real robotic experiments. Beyond demonstrating improved containment and more efficient guidance, the results highlight the method’s ability to maintain order and cohesion within the group, adapt to dynamic herd configurations, respect inter-individual spacing, and achieve guidance in less time. These advantages position our strategy as a highly effective and scalable solution for advanced robotic herding tasks.