arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2212.12554eess.SYcs.SY
收藏

具有避障和规模缩放的多智能体系统的动态循环编队:一种群集方法

Dynamic Circular Formation Of Multi-Agent Systems With Obstacle Avoidance And Size Scaling: A Flocking Approach

Seyed Mohammad Mahdi Seyed Sajadi, Hajar Atrianfar

摘要:

采用群集方法的编队控制是一种有效的方法,它可以在不确定智能体位置的情况下到达编队。针对二阶多智能体系统,研究了如何以特定的几何模式围绕领导者或目标形成圆形队形。这意味着多边形的形成是由任意的初始条件形成的。为了创建循环的形成,两个潜在的功能项已被用于代理和领导者代理的相互作用。在我们的方法中,如果一些故障发生在圆形的形成和一些代理失败,正多边形的形成仍然会形成较少的代理。保证了单圈编队和无碰撞运动的避障。为了更好地机动和通过障碍物,提出了一种具有尺寸缩放的圆形编队。此外,几个圆所需的半径可以达到与代理领导人的潜在功能的变化。在这项工作中,优化算法与不同的情况下进行比较,计算我们的算法的参数。

英文摘要:

Formation control with the flocking approach is an efficient method that can reach the formation without determining the agent's position. This paper focuses on reaching the circular formation around the leader or target with a specific geometric pattern for the second-order multi-agent system. This means that the polygon formation is formed with arbitrary initial conditions. To create the circular formation, two potential function terms have been used for agent-agent and leader-agent interaction. In our approach, if some faults occur during the circular formation and some agents fail, the regular polygon formation will still form with fewer agents. Obstacle avoidance for a single-circle formation and collision-free motion is guaranteed. A circular formation with size scaling is proposed to better maneuver and pass through obstacles. Also, several circles with the desired radius can be reached with changes in the agent-leader potential function. In this work, optimization algorithms with different scenarios are compared to calculate the parameters of our algorithm.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑