Research Of Inertially Stabilized Platforms Based On Virtual Mems Gyro Technology

A MEMS gyroscope as a kind of potential inertial components measuring angularvelocity, has no high-speed rotor, no mechanical wear, and no mass imbalancephenomenons compared with traditional gyros, but has an advantage of small size, lowcost and low power compare with optical fiber and laser gyros. However, lowmeasurement precision and large noise of practical MEMS gyros gives rise to limitationof high precision applications. It’s in such a research background that we study how toobtain high-precision sensitive angular velocity employing low-cost MEMS gyros.Besides, MEMS gyros using as the feedback component of platforms, the research issueis extended in turn about the modeling and analysis of ISPs(Inertially StabilizedPlatforms) with two axises and four gimbals,and stabilization control.A kind of configuration called virtual gyro system is based on array gyros from theperspective of reliability. The actual physical implementation of the system need notconcern us here, but some theoretical results are in order. Firstly, Allan variance isapplied to model random error for the single gyro and also the Allan variance statistic asa random variable is derived. Secondly, the key parameters about the gyro, which areimportant in the following data fusion, are identified from the simulation data using theGauss-Markov theorem verifying the correctness of Allan variance statistic. Lastly, aoptimal Kalman filter is designed and carried out to fuse data from the array gyros asthe subsystem. Implementing the virtual dynamic filter in Matlab/Simulink software,there results a good performance having a higher precision of angular velocity to someextent, which proves the validity of the method.On the other hand, the kinematics and dynamics model is established and results inthe control system block of the multi-axis gimbals. The inner stabilization loop is ourkey concern, focusing on stick-slip motion and its reasons causing the LOS ‘shaking’.On the basis of conventional PI controller,additional compensator MRAC is added tocompensate the nonlinear friction,here as a coulomb friction. A combination of PIcontroller and friction compensator shows that it’s not only effective for the nonlinearfriction, but also has the ability of disturbance rejection and the adaptability for systemparameter changes.