Application of an Improved Lifting Wavelet Method in the GNSS/INS Integrated Navigation Receiver

The noise and bias instability of low-cost MEMS sensors can cause serious location errors in the GNSS/INS integrated navigation system which limit the accuracy and errors in terms of position, velocity and attitude grow rapidly in stand-alone mode. In order to satisfy the de-noising requirements of MEMS sensors for the GNSS/INS integrated navigation receivers that have strict restrictions in size and cost, an improved lifting wavelet method is proposed which is modified to fit the scenario of real-time response in practical embedded platforms. Firstly, we analyzed the error model of the MEMS sensors and introduced Allan variance method. Secondly, principle of an improved lifting wavelet transform algorithm is introduced. We modified the algorithm by recursive threshold selection which made it possible to work in the embedded system. The algorithm is coded and compiled in C language and stationary simulation and Allan variance results verified the effectiveness of the algorithm. Finally, we successfully transplanted the algorithm into the software of a self-developed strap-down GNSS/INS integrated navigation receiver. The experiment results indicate that the method is beneficial to the improvement of the accuracy and stability of the integrated navigation system by effectively reducing the random noise of the low-cost MEMS sensors.