Sins Error Analysis And Algorithm Design

With the rapid development of electrical technology and computer technology, the strap-down inertial navigation system (SINS) has superseded the platform inertial system gradually and represents the future developing direction of the inertial technology. Nowadays, the accuracy level of SINS in our country is still low in our country. To solve this problem, one way is design high precision inertial instruments (gyro and accelerometer), but this method is limited by the material and technologic level. Another way is design effective and high precision strap-down algorithms. The attitude algorithms play important roles in strap-down system algorithms. Many scholars do a lot of works on optimal and high precision attitude algorithms, and get lots of outcome, and the research of this aspect is going on. This paper will do some research on high precision attitude algorithms on the base of the predecessors' production. The following works have been done in this paper:1. The research track of traditional attitude algorithms is analyzed. The traditional algorithms are also packed up.2. A new research track is proposed. The method of improving algorithms precision will not limited to adding the number of gyro output sampling in a rotation vector update cycle. And higher order items are added in rotation vector expression. New style of algorithm is designed.3. The coefficients of two sampling and three sampling algorithms are educed on polynomial motion model. The predigested form of three sampling algorithms is also concluded.4. Another group of coefficients of two sampling and three sampling algorithms are educed on coning motion model. The algorithms generalization is researched. And algorithms excursion is computed and compared.5. The simulation of rotation vector algorithms in different environment are given. The performance of rotation vector algorithms is compared by attitude errors. And the results indicate that the new algorithms can improve precision on precondition of not add a lot computation.