| Under the background of rapid development of infrastructure construction in China,Intelligent city,Internet of things,Location-based service and other emerging technologies are in the ascendant,smart phones stand out in the market and social life,and become an indispensible part of modern human daily activities.From observing stars,viewing terrain,drawing simple maps to satellite navigation,inertial navigation,indoor positioning,navigation and positioning technology has been updated many times,along with the progress of human society has made great progress.The acquisition of location information is a bright spot of smartphones,which meets the important needs of human for navigation and positioning.Android system occupies a huge market supports the output GNSS raw measurements,which provides a platform for studying the GNSS location algorithm of smartphones.This paper tests the localization performance of various localization algorithms on Android mobile phones.In view of the current hardware constraints of mobile phones for carrier phase,mobile phone usage scenarios are often single-point positioning,as well as cell phone power saving design criteria and other prerequisites,non-differential single-point positioning algorithm and self-contained sensors fusion algorithm are undoubtedly the most implementable and the most suitable application scenarios.This paper introduces the principle and method of obtaining the GNSS raw measurement in Android system,designs static and dynamic experiments,and tests the positioning accuracy of many GNSS algorithm,differential positioning and precise single point positioning,on the Android mobile phone.Then,the key steps based on PDR algorithm in Android mobile phone are introduced in detail to lav the foundation for self-contained sensors fusion algorithm.Through comparative analysis of previous studies,the mathematical model of PDR key steps is selected as peak detection method,linear step estimation model,accelerometer and magnetometer navigation estimation,and the feasibility and reliability of the selected model are verified experimentally.Finally,based on the aforementioned GNSS and PDR algorithms,the GNSS and PDR fusion localization methods are proposed.Considering the nonlinear characteristics of the system equations,the extended kalman filter is selected to fuse the GNSS location and PDR steplength and heading.Similar to the loose combination concept in traditional inertial navigation,the fusion algorithm has shallow coupling degree,simple and easy to implement.Filtering GNSS drift points,correcting the accumulated error of the PDR algorithm,can still provide reliable positioning for a short time when the satellite signal is blocked.The paper has 52 pictures,11 tables and 81 references. |
