Distributed and Differentiable Vector Field Control Within a Curved Virtual Tube for a Robotic Swarm Under Field-of-View Constraints

• 成果类型: Article
• 署名作者: Gao, Yan; Bai, Chenggang; Quan, Quan
• 署名单位: Beihang University; Tiangong University; Tianmushan Lab
• 刊物名称: IEEE TRANSACTIONS ON AUTOMATIC CONTROL
• ISSN/ISSBN: 0018-9286
• DOI: 10.1109/TAC.2025.3535755
• 发表日期: 2025
• 页码: 5190-5205
• 关键词: robots; Robot kinematics; vectors; Collision avoidance; safety; PLANNING; COSTS; Trajectory planning; switches; mathematical models; Field-of-view (FOV) constraint; passing-through control; robotic swarm; vector field; virtual tube

摘要

To guide a robotic swarm in a cluttered environment, a curved virtual tube is proposed in our previous work. In this article, the swarm controller design is generalized to the condition that all robots have field-of-view (FOV) constraints. Two distributed vector field controllers are proposed for different objectives. One controller focuses on the robots' safety in theory, while the other focuses on the feasibility in real practice at the cost of sacrificing safety slightly. Different from existing results, no centralized planning or switching mechanism is needed as robots are controlled to move forward within the curved virtual tube. To guarantee that these two controllers are always twice differentiable, we model the curved virtual tube in the Frenet-Serret frame using appropriate mathematical descriptions. Finally, the effectiveness of the proposed method is validated by numerical simulations and real experiments. We release our implementation as an open-source package for the community.