Research Of Gravity Error Compensation In Long-term High Accuracy Inertial Navigation System

The self-contained nature of inertial-navigation systems and their immunity to enemy jamming make them extremely attractive. Long-term high accuracy inertial navigation depends on not only the quality of the inertial sensors (accelerometers and gyros),but also the ability to compensate for the gravity disturbances. Gravity disturbances are the errors between true gravity and the gravity calculated by gravity model,and gravity compensation is the compensation for gravity disturbances. During short-term low accuracy INS, substituting model gravity obtained from certain gravity model approximately for true gravity is adequate enough. However, during long-term high accuracy INS, the east position error caused by the neglect of the vertical channel gravity disturbance will grow unboundedly with time; With the development of high accuracy INS,the precision of accelerometers and gyroscopes are highly enhanced, the error of gravity becomes the main error of INS. This paper talks about two ways of gravity compensation:"Linear direct interpolation"(DLI) and"Least square collocation"(LSC). And the main contributions include the following aspects:1. Analyze the position error affected by gravity disturbances on INS. According to navigation error model, we can extract error model of single channel, base on the model, we can deduce the formula between the north position error and north gravity disturbance. Formula is validated through the theory of"deflections of vertical", the analysis of simulation and experiment data.2. Introduce several kinds of gravity models and several ways to measure gravity disturbances. The gravity model used for simulation in this paper is WGS84, and"Least square collocation"compensation method use reciprocal distance model(RDM) as the gravity disturbance model. For the sake of obtaining gravity data to validate the effect of LSC and DLI, this paper talks about how to measure gravity disturbances in several ways, and focus on the comparison among them.3. Research the relationship between horizontal position error and gravity disturbance of different channels. The result concludes that: the horizontal position errors grow in proportion with gravity disturbances,but the error won't grow unboundedly with time; however the horizontal position error caused by gravity disturbances of horizontal channel is too large to be neglected so that we need to compensate them .4. Deep research is carried out aiming at the comparison of two compensation method. After introduce the principle of LSC and DLI, we analyze their resistance to noise and research the effect of density of the gravity data. The result shows that: both of them are good at resisting noise, and the larger the density of the data, the better compensation will be. And for most of time LSC is better than DLI. At last we use experiment to validate the result concluded by simulation.