Onboard multi-rate tube model predictive position control for a hexacopter: From design to outdoor flight experiments
Onboard multi-rate tube model predictive position control for a hexacopter: From design to outdoor flight experiments
复制标题
DOI：
10.1109/icat.2017.8171596
复制
发表时间：
2017-10
期刊：
影响因子：
--
通讯作者：
P. Niermeyer;Markus Herb;B. Lohmann
中科院分区：
文献类型：
--
作者：
P. Niermeyer;Markus Herb;B. Lohmann
模型预测控制（MPC）是一种主动控制受限对象的备选方案。在管模型预测控制（TMPC）中，最优性被鲁棒性所取代，鲁棒性是飞行控制中的一个关键方面，在飞行控制中，被控对象受到许多干扰和不确定性，尤其是与空气动力学和风有关的干扰和不确定性。然而，MPC带来了大量的计算量，从而降低了控制律的执行率，从而加剧了干扰抑制。这个问题可以通过以不同的速率执行基于优化的前馈和简单的反馈控制来解决。因此，作者提出了一套完整的多速率控制设计、软硬件实现、测试和飞行实验进行的工作流程。首先，讨论了控制律。接下来，介绍了定制的板载控制软件，并在定制的硬件在环（HIL）仿真中进行了基准测试。最后，飞行实验证明了TMPC的性能以及该方法所带来的设计、实现和测试的便利性。
Model Predictive Control (MPC) is a candidate solution to aggressively control a constrained plant. In Tube Model Predictive Control (TMPC), optimality is traded for robustness which is a key aspect in flight controls where the plant is subject to many disturbances and uncertainties especially related to aerodynamics and wind. Yet MPC comes with large computational effort and thus reduced execution rates of the control law which aggravates disturbance rejection. This issue can be addressed by executing optimization-based feedforward and simple feedback control at different rates. Therefore, the authors present a complete workflow for multi-rate control design, soft- and hardware implementation, testing and flight experiment conduction. First, the control laws are discussed. Next, the custom onboard control software is presented and benchmarked in a tailored Hardware-in-the-Loop (HIL) simulation. Finally, flight experiments demonstrate the performance of the TMPC and the ease of design, implementation and testing entailed by the presented approach.