Experimental Study on GPS Non-linear Least Squares Positioning Algorithm

The paper gives an overview of non-linear least square method, and presents GPS satellite pseudorange observation equations, which are usually over-determined because there are only 4 unknowns but normally more than 4 satellites being tracked, resulting in more than 4 pseudorange observation equations. Non-linear least squares(NLS) algorithm to solve pseudorange observation equations in order to find position and clock bias solutions are described in detail. There are three tests designed, one vehicle field test and two simulation tests. There are three NovAtel made GPS receivers being used in vehicle field test, while GSS6560 GPS satellite signal generator being used in simulation tests. By employing a set of differential GPS receivers, it is possible to establish rover station receiver to give a relatively high accurate reference to be compared by NLS and another GPS receiver in the vehicle field test. Results from NLS demonstrate that NLS can achieve high accuracy. Experimental study also shows that NLS algorithm possesses similar accuracy level to that of NovAtel Receiver, which represents world leading level.

[1]  Robert M. Rogers,et al.  Applied Mathematics in Integrated Navigation Systems , 2000 .

[2]  Peter Teunissen,et al.  GPS for geodesy , 1996 .

[3]  Xianlin Huang,et al.  Geomagnetism GPS SINS Integrated Navigation Based on an Improved Square-Root UKF , 2009, 2009 Second International Conference on Intelligent Computation Technology and Automation.

[4]  Bradford W. Parkinson,et al.  Global Positioning System , 1995 .

[5]  Xiaojing Du,et al.  Position accuracy evaluation of GPS receiver under small UAV flight environment , 2009, 2009 9th International Conference on Electronic Measurement & Instruments.