Optimal trajectory generation for time-to-contact based aerial robotic perching
Optimal trajectory generation for time-to-contact based aerial robotic perching
复制标题
DOI：
10.1088/1748-3190/aaeb13
复制
发表时间：
2018-11
影响因子：
3.4
通讯作者：
Haijie Zhang;Bo Cheng;Jianguo Zhao
中科院分区：
文献类型：
--
作者：
Haijie Zhang;Bo Cheng;Jianguo Zhao
许多生物有机体（例如昆虫、鸟类和哺乳动物）依赖于对称为接触时间（TTC）的信息变量的感知来控制它们的运动以执行各种任务，例如避开障碍物、着陆或拦截。TTC被定义为在保持当前速度的情况下接触物体所需的时间，最近已被用于各种任务中的机器人运动控制。然而，大多数现有的机器人应用的TTC简单地控制TTC是恒定的或不断减少，没有充分探索TTC的适用性。在本文中，我们提出了两个阶段的TTC为基础的战略，并将其应用于空中机器人栖息。与所提出的策略，我们可以生成参考轨迹TTC实现非零接触速度所需的栖息，这是不可能的恒定或不断减少TTC策略，但鲁棒栖息性能至关重要。我们进行了模拟，以验证所提出的策略在更短的时间栖息和满足更多的约束方面的优越性。此外，适当设计的控制器，我们进行实验上的手掌大小的四轴飞行器跟踪计划的参考轨迹，实现空中机器人栖息。本文的研究可以很容易地应用到飞行机器人的控制与视觉反馈的栖息，并可以启发更多的替代形式的TTC为基础的规划和控制的机器人应用。
Many biological organisms (e.g. insects, birds, and mammals) rely on the perception of an informational variable called time-to-contact (TTC) to control their motion for various tasks such as avoiding obstacles, landing, or interception. TTC, defined as the required time to contact an object if the current velocity is maintained, has been recently leveraged for robot motion control in various tasks. However, most existing robotic applications of TTC simply control the TTC to be constant or constantly decreasing, without fully exploring the applicability for TTC. In this paper, we propose two-stage TTC based strategies and apply them to aerial robotic perching. With the proposed strategies, we can generate reference trajectories for TTC to realize the non-zero contact velocity required by perching, which is impossible for constant or constantly decreasing TTC strategy, but of critical importance for robust perching performance. We conduct simulations to verify the superiority of the proposed strategies in terms of shorter time for perching and satisfying more constraints. Further, with properly designed controllers, we perform experiments on a palm-size quadcopter to track the planned reference trajectories and realize aerial robotic perching. The research presented in this paper can be readily applied to the control of flying robot for perching with visual feedback, and can inspire more alternative forms of TTC based planning and control for robotic applications.