Experiment Videos of a Force-Based Organization and Control Scheme for the Non-Prehensile Cooperative Transportation of Objects

The videos provided show experimental results of the cooperative object transportation using mobile robots with on-board force control. In particular, the mobile robots shall transport four different polygonal, but non-convex, objects along predefined paths. The distributed formation synthesis explicitly takes the robots' maximum pushing force into account such that a closed-set manipulation space in terms of a zonotope follows. This zonotopic manipulation space, following from the Minkoswki sum of the individual manipulation capabilities of the robots, is visualized in the fifth video (Visualization_Fig4a.mp4) using data from the corresponding hardware experiment (Rectangle_N4_square.mp4). Novelly, a lightweight quadratic program runs on each robot and determines in a decentralized manner the desirable individual pushing forces suitable to transport the object. These pushing forces are then governed by means of hybrid position-force controllers running at 100 Hz. As for formation finding, no central entity is used for control purposes. The tasks are accomplished in a purely distributed manner using inter-robot communication. For each robot, the pushing force is measured using a self-designed force-sensing unit mounted on board of each mobile robot. In the videos, the direction of the uniaxial and unilateral force sensor is indicated by white rectangles. Moreover, the measured force is visualized by red rectangles superimposed onto the white ones.

所提供的视频展示了利用车载力控制技术的移动机器人进行协作式物体运输的实验结果。特别是，移动机器人需沿预定义路径运输四种不同形状的多边形，但非凸形的物体。分布式编队合成明确考虑了机器人的最大推力，从而确保了基于区间的闭合操作空间。这一基于机器人各自操作能力的 Minkowski 和的区向量操作空间，在第五个视频中（Visualization_Fig4a.mp4）通过对应硬件实验（Rectangle_N4_square.mp4）的数据进行了可视化。创新性地，每个机器人上运行一个轻量级的二次规划程序，以去中心化的方式确定适合运输物体的理想推力。这些推力随后通过运行在100 Hz频率的混合位置-力控制器进行调控。至于编队定位，控制过程中未使用任何中心实体。任务以纯分布式方式通过机器人间通信完成。对于每个机器人，使用安装在移动机器人上的自设计力感应单元来测量推力。在视频中，单轴单向力传感器的方向由白色矩形指示。此外，通过红色矩形叠加在白色矩形上，对测量的力进行了可视化。