| The inertial navigation system (INS) determines vehicle’s position and attitude informationonly by the calculation of navigation computer, which works by the outputs of inertial sensors.This characteristic of INS makes it not sensitive to disturbances of outer environment anddifficult to be found, thus it is more suitable for bad weather condition and battlefieldenvironment. SINS has become a research focus in inertial navigation engineering field theseyears because of its smaller size, lower cost and high alignment accuracy. Based on theapplication of shipboard SINS, coarse alignment is researched in this paper. The INS soloworking characteristic will result in the weakness of error accumulation, therefore optimizingthe initial alignment method has important theoretical and practical values for improving theaccuracy of whole system.The related background is explained firstly in this paper, and then SINS is introduced indetail. Basic equations of INS are given and a system block diagram of SINS is set up. Becauseattitude updating algorithm is the most important part of SINS, optimized three-samplealgorithm is deduced. The errors of SINS are given too, which are significant to latersimulation.Taking initial alignment process of ship’s SINS as the research object, analytical coarsealignment, horizon leveling plus azimuth estimation coarse alignment and inertial frame coarsealignment based on attitude matrix are compared in this paper. Firstly, equations of three coarsealignment methods are inferred. Then, methods are tested on stationary and swaying base,simulation results are given in the end.As quaternion has easier updating equation and faster computing speed than Euler angles,the inertial frame alignment based on quaternion algorithm is deduced in this paper. Wahba’sfunction is used to derive the optimal least-squares solutions. In addition, the method errors areanalyzed and covariance of attitude error is given. After that, two sampling methods arecompared, and a better sampling method is proved.Errors of accelerometer could be reduced through integration process in the algorithmproposed; however, high-frequency disturbances and errors of gyro cannot be filtered.According to these shortages, a Hidden Markov Model Kalman Filter (HMMKF) is presentedin this paper to pre-filter the noise of inertial sensors. The proposed inertial frame alignmentmethod based on quaternion is improved and the simulation results are given. |
