CAREER: Advanced Control Algorithms for Active Materials Actuators Used in Nanoscale Positioning

对于纳米级的操作，高度精确的定位能力是一个非常基本但不可或缺的要求。这一领域可能的应用范围从分子结构的组装到细胞操作和高精度的加工任务。目前使用的纳米操纵设备通常依赖于使用来自活性材料的致动器，主要是压电体和磁致伸缩材料。只有通过这些材料及其固有的特性，才有可能在极小的体积内实现有效的驱动。另一方面，这些材料表现出高度的非线性行为，其特征是出现磁滞回线。这极大地影响了它们的控制性能，并限制了它们在常规(PID)反馈回路下的低带宽和低冲程应用的适用性。对于几乎每一项纳米制造任务来说，在保持高精度的同时提高速度和效率具有重要意义。处理这个问题的一种方法是在控制算法中加入材料行为的模型。最近，PI在形状记忆合金领域开发的一种新方法为一种极其高效的算法奠定了基础。这种方法不仅扩展了当前使用的方法的范围，以考虑调整速度或能量消耗等优化标准；由于其计算速度快，它显然具有实时能力。在微观尺度上，形状记忆合金的物理机制与压电材料和磁致伸缩材料的物理机制密切相关，基于这一观察结果，将为这三种材料建立一个统一的模型。基于这种统一的模型，将实时最优控制方法扩展到考虑闭环反馈和参数更新，并将在实时硬件环境中进行验证。该教育计划涉及四项举措，将促进从本科生到博士水平的新的多学科课程的发展，以及积极的本科生研究整合。第三项努力将增加学生接触国际问题的机会，并通过提供与德国领先的纳米技术机构之一合作的机会，提高他们的外语熟练程度。最后一项倡议是开发基于网络的课程材料，包括通过基于活动材料模型的Java版本的模拟来进行“虚拟实验室实验”。

For manipulations at the nanoscale, highly precise positioning capability is a very basic, yet indispensable requirement. Possible applications in this area range from the assembly of molecular structures to cell manipulations and highly precise machining tasks. Currently employed devices for nanomanipulation commonly rely on the use of actuators from active materials, mostly piezoelectrics and magnetostrictives. It is only through these materials and their inherent properties that the possibility of efficient actuation in an extremely small volume is offered at all. On the other hand, these materials exhibit a highly non-linear behavior characterized by the occurrence of hysteresis loops. This greatly afflicts their control performance and limits their applicability to low bandwidth and low stroke applications under conventional (PID) feedback loops. It is of great relevance for almost every nanomanufacturing task to improve speed and efficiency while at the same time maintaining high precision. A way to treat the problem is to incorporate a model of the material behavior into the control algorithm. A novel approach in the field of shape memory alloys developed by the PI has recently established the foundation for an extremely efficient type of algorithm. This approach not only extends the range of currently employed methods to account for optimality criteria like speed of adjustment or energy consumption; it is clearly real-time capable due to its high computational speed. Based on the observation that, on the icroscale, the physical mechanisms in shape memory alloys are closely related to the ones observed in piezoelectric and magnetostrictive materials, a unified model will be developed for all three materials. Based on such a unified model, the real-time optimal control approach will be extended to account for closed-loop feedback and parameter updating, and it will be validated in a real-time hardware environment. The educational plan involves four initiatives and will promote the development of new multi-disciplinary courses from undergraduate to Ph.D. level, as well as active undergraduate student research integration. The third effort will increase the exposure of students to international issues and increase their foreign language proficiency by offering possibilities to work with one of Germany's leading institutions in nanotechnology. The final initiative is the development of web based course materials, including "virtual lab experiments" through a simulation based on JAVA versions for active materials models.