Suppression Methods On Predominant Errors Of The MIMU/BDS Deeply Integrated Navigation System

Behaving complementary error characteristics, INS/GNSS integrated system is becoming a key navigation equipment for the aerocraft or vehicle. According to data fusion strategies, three INS/GNSS integration architectures are classified, namely loose integration, tight integration and deep integration. For the deep integration system, improved navigation precision, high dynamic adaptability and anti-jamming capability can be gained as INS and GNSS aid each other. This study is carried out based on our built domestic MIMU/BDS deeply integrated navigation system, focusing on predominant errors and their suppression methods of the system under the high dynamic environment. Main conclusions are as follows:(1) A novel method to suppress the temperature drift error of the sillicon micromachined gyroscope is proposed and studied. Different from traditional temperature compensation technique, the main idea of this method is minishing or eliminating the demodulation-incoordination-angle, which is a key parameter in the transfer chain of the temperature drift error. The generation mechanism and temperature sensitivity of the demodulation-incoordination-angle are analyzed, then a high precision frequency tracking loop is designed to eliminate demodulation-incoordination-angle. Further, Kalman filtering technique is adoptd to estimate the angular rate and demodulation-incoordination-angle instead of traditional synchronization demodulator. By virtue of the method, the error caused by the demodulation-incoordination-angle is separated from the gyroscope output, thus the temperature sensitivity of the gyroscope is effectively depressed. Considering unknown and changing noise characteristic, adaptive and strong tracking Kalman filtering technique is also introduced to enhance the robustness. The proposed method is validated by temperature experiments. When the temperature changes by 100℃, the bias drift is up to 4.5950o/s for the traditional synchronization demodulator, while 0.5086o/s for the proposed method. The experimental results show that nearly 10 times improvement on the temperature stability is gained.(2) The relation between the performance of the receiver carrier tracking loop with MIMU aiding and the precision of MIMU measurement is investigated. A formula to describe the relation is deduced, by which the precision floor of the MIMU to maintain carrier loop lock is given. The analyses and simulation results show that the bias instability of the accelerometer should be less than 29.72 mg, and 490.44°/h for the gyroscope when carrier loop is designed as three order PLL with 18 Hz noise bandwidth in high dynamic entironment.(3) To reduce the carrier loop error due to inaccurate MIMU measurement, a novel aiding scheme based extended measurement kalman filter(EMKF) is proposed. In the EMKF method, the aiding information is used as a measurement of the Kalman filter, which is different from the tradional idea where the aiding information is added with the estimated doppler from the tracking loop. For low precision MIMU, the EMKF method can reduce the effects of aiding errors due to MIMU inaccuracy on the tracking loop, hence an improvement on the tracking performance. The superiority of the EMKF method is testified by both theory and simulation analysis. In the high dynamic environment, the loop will lose of lock when the bias instabilitis of MIMU is up to 29.6mg and 296o/h for the tradtional method, but 76.2mg and 762o/h for the proposed EMKF method.(4) A heading angle determination method utilizing acceleration information derived by the receiver is put forward and is used for restraining the heading angle error of the integration navigation system hence an improvement on the heading angle accuracy of the system. In the method, the heading angle is calculated using the carrier’s acceleration, and position, velocity, attitude angles from the integration system and the specific force sensed by the accelermeters, which is independent of the output of gyroscopes. When the obtained heading angle is used as a measurement of the integration filter, refering to the INS/GNSS/magnetometer integrated system, the observability of the heading angle error can be improved. The proposed method is validated by both semi-physical simulations based on the imu/gnss signal generator and experiments. The road tests results show that the standard deviation of the heading angle error is reduced from 2.53° to 0.45°.