MRI: Acquisition of dSPACE Real-Time Development System and Electromechanical Apparatuses for Research in Smart Control Systems

MRI：收购 dSPACE 实时开发系统和机电设备，用于智能控制系统研究

• 批准号: 0320867
• 负责人: Marc Karam
• 金额: $ 7.23万
• 依托单位: Tuskegee University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2003
• 资助国家: 美国
• 起止时间: 2003-08-01 至 2005-07-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=0320867&HistoricalAwards=false
• 关键词: MRI; Acquisition; dSPACE; Real; Time

In the department of Electrical Engineering at Tuskegee University, the PI and three students(two graduates and one undergraduate) are conducting research in the area of Smart ControlSystems. The following two projects are underway: robust optimal control using a recurrentdynamic neural network, and compensation of global feedback linearization control of nonlinearsystems using neural networks. However, these research activities are confined to theoreticalinvestigations and applications to computer simulated system models using Matlab/Simulinksoftware from the MathWorks Inc. No physical plants and corresponding real-timeimplementation software are currently available to test the smart control algorithms and refinethem.The requested instrumentation to be acquired includes the ACE1103PX4CLP prototypingsystem from dSPACE Inc., which contains powerful hardware and comprehensive software toolsfor development of real-time control systems. In order to apply the dSPACE products to actualplants, we are in addition requesting the acquisition of the following electromechanicalapparatuses from ECP, a leading manufacturer of mechanisms for research in control: theRectilinear Mass/Spring/Damper (ECP Model 210), the Industrial Plant Emulator (ECP Model220), the Inverted Pendulum (ECP Model 505), the Magnetic Levitator (ECP Model 730), andthe Control Moment Gyroscope (ECP Model 750).The acquired instrumentation will contribute to greatly advancing the above-listed researchprojects by providing the possibility of implementing the current Matlab/Simulink codes andmodels of the neural controllers on dSPACE hardware, and testing them in real-time as appliedto actual electromechanical apparatuses, as well as checking the robustness of the neuralcontrollers in the presence of noise and disturbances. The feedback obtained from the actualimplementation of the various neural algorithms will contribute to significantly evaluating,refining, and improving the smart controllers design.The intellectual merit of this proposal resides in the fact that, if funded, the smart controlresearch at Tuskegee University would be greatly enhanced, which in turn would affect theglobal research community in the area of artificial intelligence and control, since several journaland conference publications have resulted thus far from this research, and more would resultwith higher quality if the instrumentation is acquired.Broader impacts of the requested instrumentation include introduction of a new graduatecourse in Advanced Control Design at Tuskegee University, which is an HBCU committed to thehigh quality education of African-American students. Moreover, the acquired equipment willboost the research level and researchers motivation and learning experience and contribute toattracting new graduate students to Tuskegee University that are interested in the area of smartcontrol systems. The development of this area that is expected to occur due to the requestedinstrumentation will have the additional impact of attracting research funding from industry andgovernment agencies. Furthermore, the experience that students will have with dSPACEproducts will be very beneficial for their future career, since these products are used in majorindustries worldwide. Finally the acquired equipment will encourage the multidisciplinary trendin the College of Architecture, Engineering, and Physical Sciences at Tuskegee University bybeing made equally available to the faculty and students of the Aerospace, Chemical, Electrical,and Mechanical Engineering departments, which all offer courses in control systems.

在塔斯基吉大学电气工程系，PI和三名学生（两名研究生和一名本科生）正在进行智能控制系统领域的研究。以下两个项目正在进行中：使用递归动态神经网络的鲁棒最优控制，以及使用神经网络补偿非线性系统的全局反馈线性化控制。然而，这些研究活动仅限于理论研究和使用MathWorks公司的Matlab/ simulink软件对计算机模拟系统模型的应用。目前还没有实体工厂和相应的实时实现软件来测试智能控制算法并对其进行改进。所要求的仪器包括来自dSPACE公司的ACE1103PX4CLP原型系统，该系统包含强大的硬件和全面的软件工具，用于开发实时控制系统。为了将dSPACE产品应用于实际工厂，我们还要求从ECP（一家领先的控制研究机构制造商）购买以下机电设备：直线质量/弹簧/阻尼器（ECP 210型），工业工厂模拟器（ECP 220型），倒立摆（ECP 505型），磁悬浮（ECP 730型）和控制力矩陀螺仪（ECP 750型）。通过提供在dSPACE硬件上实现当前Matlab/Simulink神经控制器代码和模型的可能性，以及在实际机电设备上对其进行实时测试，以及检查神经控制器在存在噪声和干扰时的鲁棒性，所获得的仪器将有助于大大推进上述研究项目。从各种神经算法的实际实现中获得的反馈将有助于显着评估，改进和改进智能控制器设计。该提案的智力价值在于，如果得到资助，塔斯基吉大学的智能控制研究将大大加强，这反过来将影响人工智能和控制领域的全球研究界，因为一些期刊和会议出版物迄今为止已经产生了这项研究，如果获得仪器，更多的结果将具有更高的质量。所要求的仪器的更广泛影响包括在塔斯基吉大学引入一门新的高级控制设计研究生课程，这是一所致力于非洲裔美国学生高质量教育的HBCU。此外，获得的设备将提高研究水平和研究人员的动机和学习经验，并有助于吸引对智能控制系统领域感兴趣的新研究生到塔斯基吉大学。由于所要求的仪器，预计该领域的发展将对吸引工业和政府机构的研究资金产生额外的影响。此外，学生使用dspace产品的经验将对他们未来的职业生涯非常有益，因为这些产品在世界各地的主要行业中都有使用。最后，获得的设备将通过向航空航天、化学、电气和机械工程系的教师和学生平等提供，鼓励塔斯基吉大学建筑、工程和物理科学学院的多学科趋势，这些系都提供控制系统课程。