Research On Algorithm Rapidity And Errors Calibrated Methods Of The Rate Biased RLG North-finder System

The rate biased RLG north-finder system plays an important role in the civil and military areas as it can provide the local vertical plumb and orientation benchmark. It can work in formidable natural conditions and has high precision. Precision and rapidity are important technical targets of the north-finder system. On the basis of the errors characteristics of the rate biased RLG north-finder system, this paper focuses on improving the norhfinding precision and rapidity by using the error parameters calibration and the algorithm optimization. Main researches are as the follows.1. Firstly, different errors influence to the precision is analyzed, including the IMU installation errors, RLG scale factor errors, RLG biase errors, accelerometer biase errors, and base disturbance errors. Secondly, it is proposed that installation matrixes of RLGs and accelerometers are calibrated at the mechanical dither biased mode of RLGs. And frames transfer theory and RLGs'rake angles calibrated theory are presented. The angle between the prism datum line and the geographic north can be gained continuously at the rotating condition, so the system can output the northfinding result quickly. Thirdly, a calibration technique for the rate biased RLGs'scale factors and biases are realized.2. A coarse align algorithm on the rotating base is realized, and experimental results show that coarse alignment algorithm has high precision. Because the attitude angles'errors are smaller, convergence time of Kalman filter is shortened. Therefore, the northfinding rapidity is improved.3. According to the errors characteristics of the north-finder system, and based on the observability degree analysis and mathematical emulation, the new northfinding algorithms are proposed as follows: firstly, based on the errors characteristics of rate biased RLGs and accelerometers in the temperature rising process, second-order nonlinear error models are established respectively to make online nonlinear estimation; secondly, dimensionality reduction Kalman filter is designed to improve the rapidity. Experimental results show that the compensation methods achieve attractable effects, and Kalman filter estimates the RLG biase errors fast.4. Multi-groups results at one position and muti-positons results show that by using the northfinding algorithm, the northfinding precision in 3 minutes is better than 1′(1σ)in laboratory environment, and rapidity and precision of the north-finder system are improved apparently. The experiments in the outside environment prove that the north-finder system works effectively, and experimental results show that the northfinding precision in 3 minutes is better than 2(′1σ) in perturbational environment.