CAREER: Microsystem Model Identification and Control

University Of California-Los Angeles, Los Angeles CA

Investigators

Abstract

This multidisciplinary research program integrates experimental, computational, and analytical methods for micro system modeling identification, and control. Micro sensors and actuators are finding a niche in inertial navigation, vehicle stabilization, and flow control applications where power consumption, weight, and cost are primary considerations. The manufacturing technology of these devices, however, does not enable precise control over some important electro-mechanical properties. Furthermore, the devices can exhibit sensitivity to environmental factors, such as temperature and pressure, so additional compensation must be employed to improve linearity, dynamic range, and stability. Real-time model identification algorithms are at the heart of this capability. The algorithms enable higher levels of performance, autonomy, and reliability and can lower system costs through increased yields and increased self-testing. Increased yields for micro systems arise from the ability to self-identify and compensate for micro-fabrication errors that would otherwise limit performance or cause test failure. These algorithms, based on adaptive lattice filters, have been used for on-line calibration of several micro gyroscope prototypes. The results are extremely encouraging and new classes of model structures that incorporate nonlinearities present in micro sensors have been developed. Complementary to this modeling effort are the development of novel nonlinear control architectures and synthesis methods that address the unique performance objectives of micro systems. A number of devices, including micro-flow control actuators and other sensor prototypes, are being tested with these new tools. Education is enhanced through the undergraduate Sensor and Actuator Laboratory. This laboratory introduces students to industrial sensor and actuator instrumentation and provides them with research experience by exploring many of the problems associated with micro system design, calibration, and control.