RTK/Accelerometer Tightly Coupled Deformation Monitoring Algorithm based on Baseline Shift Virtual Observation Constraint

The precision of deformation monitoring with Global Navigation Satellite System (GNSS) technology will degrade in complex environments, such as landslides. Furthermore, the loosely coupled (LC) model based on GNSS and accelerometer offers only limited improvement when dealing with continuous unreliable outliers. In this study, our work concentrates on the multi-sensor data process, and we introduce a tightly coupled (TC) algorithm that combines real time kinematic (RTK) and accelerometer, in which the state and observation equations are described in detail. To address the issue of the accelerometer baseline shift (BS) potentially absorbing the errors from float coordinates when the fix ambiguities cannot be resolved, we further propose a virtual observation constraint model based on BS. This model is named as the novel tightly coupled (NTC) algorithm in this paper, aiming to slow down the error accumulation in the acceleration integral process. Two experiments were conducted: a simulated deformation experiment and a real-world landslide monitoring task, to validate the effectiveness of the proposed algorithm. The TC algorithm notably enhances the fixed rate of integer ambiguities, achieving 74.37% and 99.97% fix rate in the two experiments, respectively, which represents a 25.37% and 51.67% improvement over the RTK solution. In case of the NTC algorithm, it achieves more precise BS correction, which leads to a further positive effect on enhancing the fixed ambiguity resolution in the simulated experiment. In the real-world landslide experiment, while the improvement of fix rate is non-significant, however, the ratio value results still demonstrate a noticeable contribution.