Research On The Methods Of Airborne Gravimetry Based On SINS/DGPS

Airborne gravimetry is the method to map the regional and local gravity field on airplane, it can map the gravity field of a big area rapidly and continuously for its increased speed, greater range, significantly low cost and the potential to reach anywhere, it has been one of the key research areas in gravimetry during the recent 20 years. The beginning of airborne gravimetry is relatively late in China, and the current system can not meet the requirements of high accuracy applications. Before the national 863 major project "Developing and Integration of Airborne Gravity Survering System", this dissertation performs pre-research on airborne gravimetry based on strapdown inertial navigation system(SINS). After the deep research on the principle of the airborne gravimetry, the research lays emphases on measuring specific force and measuring vehicle acceleration. The main contributions include the following aspects:1. The principle of airborne gravimetry based on SINS is studied systematically, and the accuracy requirements for GPS and SINS are analysized according to the error model, the time and space consistency problem in airborne gravimetry is studied, and the theory for developing prototype is founded.2. For measuring specific force, the SINS/DGPS reduced order filtering algorithm has been proposed. Because the error of free SINS will accumulate along with the elapse of time, SINS/DGPS integration is needed to improve the accuracy of specific force. In traditional algorithm, Kalman filter is used to estimate the attitude error and accelerometer bias, and the estimated values are used to correct the specific force of SINS. Observability analysis shows that traditional 15-states Kalman fliter is not fully observable. So, it is considered from intuition that using the estimated values of the unobservable states to correct the specific force error of SINS should be unreasonable, but the reality is that direct compensation is feasible. Disagreement between intuition and reality means that there is inevitability has not been discovered. Through the state-transform, the system is decoupled into a fully observable and a fully unobservable subsystem, and it is found that the error of specific force can be calculated directly from the estimated values of the fully observable states, so the reduced order filtering algorithm is designed. The reduced order filtering algorithm reveals the inevitability of traditional algorithm, and makes the concept of specific force measurement clearer.3. Accelerometer is the key sensor in airborne gravimetry system, the generally used inertial grade quartz flexure accelerometer can not meet the requirement of airborne gravimetry, its accuracy is mainly affected by temperature. The three control methods of the thermal error of quartz flexure accelerometer are studied. Firstly, the crossover adjustment method which is used in the SISG of Calgary University is studied, it is showed that this method can not satisfy the requirement of practicability of project. Secondly, the modeling and compensating method of thermal error is studied aiming at the homemade quartz flexure accelerometer CHJN-2E, the static thermal error model is founded, experiments show that the accuracy is improved about a grade. And the thermal control method of quartz flexure accelerometer is studied preliminarily at the last.4. Deep research is carried out aiming at the problem of low dynamic performance of carrier phase direct method for measuring acceleration, the causation of the dynamic performance degradation was found, this discover has important value on the improvement of accuracy and resolution of acceleration. The commonly used position differentiation method needs double frequency ionospheric-free measurement, the increased noise level of ionospheric-free measurement increases the difficulty to improve the accuracy and resolution of acceleration. Because only L1 single frequency measurement is needed in carrier phase direct method, lower level of measurement noise makes it possible to improve the accuracy and resolution of acceleration. Experiment was carried out using static data with baseline length of 20km, it showed that the accuracy of carrier phase direct method was higher 34% than position differentiation method after 120 seconds low pass filtering. But the poor dynamic performance of the former algorithm makes the carrier phase direct method can not be used in practice. This dissertation performed deep research on the model of carrier phase measurement, and found that the error of model simplification is the cause of dynamic performance degradation, then a novel iteration algorithm was designed based on the characteristic of precise model to compensate the model error. This discover is verified using GPS aero dynamic data.5. The covariance of the error of the carrier phase pseudorange acceleration has been modeled, this covariance model reduces the effect of the constellation changes on the acceleration accuracy efficiently. When using least square method to calculate acceleration, the rise and set of GPS satellite and the signal interruption of low elevation GPS satellite will induce intolerant acceleration error, the reason is that least square method treats the quality of all measurement equally. If covariance can be modeled reasonably, weighted least square method can be used to reduced the constellation changes on the acceleration accuracy. This dissertation brings the covariance model which include the parameters of elevation angle, separation angle, baseline length, etc. The efficiency of the covariance model is verified using static data with baseline length of 20km, the result shows that the accuracy of vertical acceleration is improved about 44% after 200 seconds low pass filtering, and effect of the constellation changes on the acceleration accuracy is reduced efficiently; Simultaneous, its performance is better than the variance model in which the physical correlations are not taken into account.