Design For The Micro Inertial Measurement Unit Of A High Speed Underwater Vehicle And Its Related Techniques

Underwater vehicles with supercavitating and drag reduction technologies have got considerable improvement in speed and flexibility, which overcomes the defects of traditional underwater vehicles such as low speed, short voyage and bad precision, opening a wider development area for underwater superspeed vehicles. Installation of navigation equipment in such vehivles becomes essential for raising the efficiency of vehicles. Following the continuing progress in design of vehicle's structure, we have higher requirements for volume, reliability and anti-shocking. Therefore the minimization of built-in inertial navigation system becomes more urgent. Micro inertial navigation system can realize a wholly-automated navigation independent of any external information. Combining micro mechatronic technique, the micro inertial measurement unit (MIMU) possesses wide application perspective for underwater superspeed vehicles. However, comparing with conventional inertial measurement components, MIMU has lower precision, which severely handicap its application. Hence,it is significant to know how to take full advantages of microscopic inertial parts and improve the measuring precision,so that the measuring precision of MIMU can be also improved. In the paper, a series of research work was performed for improvement of reliability and precision of MIMU.First, the MIMU with nine gyroscopes was designed, in which the 4th redundant information were used by the collocation method of inclined axis redundancy to realize check and complementation for information of 3 orthogonal axes, remarkably increasing the reliability and precision of this system compared to the conventional collocation methods.Then, the feasibility of increasing the precision of the system proposed in this thesis was discussed by using a data fusion algorithm. In order to further increase system measurement precision, redundant and complementary information coming from the sensors in different locations is synthetically utilized to reduce uncertainty of measured information by use of optimally weighted least square method (OW-LSM)and finite windowing weighted algorithm(FW-LSM) and the measuring variance self-learning weighted least squares (SL-LSM) respectively.Finally, the drift characteristics of micromechanical gyroscopes were analyzed by the method of collecting a large number of sample data. The stochastic drift of micromechanical gyroscopes is more complex than systematic drift, and its modeling and compensation are very important to increasing precision of measurement information. In this thesis, certainty trend terms in drifting were abstracted based on two improved GM(1,1) models, and modeling analysis was done on the stochastic drift using AR(3) model. Allan variance analysis shows that the drift noises after compensation of GM-AR models are greatly reduced.In addition, repeated tests were done on the micromechanical gyroscopes by using the multi group data in one day and different days, showing well information repeatability of the micromechanical gyroscopes. And the stochastic drift signals were processed using Kalman filter. Although not all the stochastic drift signals were removed, they were reduced substantially. Measurement results with higher precision can be obtained in practical application by using the method mentioned above.The research results in this thesis show that the design of MIMU can make full use of the advantages of micro inertial components, increasing the reliability and precision of the system. Compensation techniques of the data fusion algorithm and drift error model can further increase the precision of micromechanical gyroscopes under the condition of existing hardware, showing a good effect.