A Tightly-Coupled GNSS RTK/INS Positioning Algorithm Based on Adaptive Lag Smoother

Abstract

How to take into account both the calculation cost and positioning accuracy when driving over a long distance in the scene of changing satellite visibility, such as urban areas and mountain roads, is a research topic worth of attention for intelligent vehicles. In this paper, a tightly-coupled RTK/INS positioning algorithm base on adaptive lag smoother is proposed. By combining the uncertainty of the state to be estimated at the current time and the quantitative score of satellite visibility, the lag-length in smoother can be adjusted adaptively, so as to ensure positioning accuracy while reducing the calculation cost of marginal benefit consumption as much as possible. The proposed algorithm is demonstrated in both the simulator and real-world urban roads. From the experimental results, it is found that the estimation accuracy achieved by the adaptive lag smoother is similar to that of smoothers with long lag length, but the time consumption is reduced by about 30%. Under the same condition that the positioning can be completed in real time, the accuracy of the algorithm in this paper is 27% higher than that of the tightly-coupled RTK/INS system based on extended Kalman filter.