Research On The Key Technology And Platform Design Of Real-time GNSS/SINS Integrated Navigation System

As the most typical positioning methods in direct positioning and dead reckoning technologies,GNSS and SINS have obvious complementary advantages.By combining them,the GNSS/SINS integrated navigation system is formed.It can provide high quality navigation services and has been utilized in weapon guidance,aviation and aerospace,mobile mapping,etc.In recent years,with the prosperity of autonomous driving,smart robot,unmanned carrier and other new technologies,the requirement for real-time,high reliability,accurate navigation in poor GNSS environment is urgent.And the performance of navigation system is becoming the key factor for those applications' promotion.As a result,this thesis focuses on key technology and platform design of real-time GNSS/SINS integrated navigation system.The main works are as follow: 1)An ARM based hardware platform for integrated navigation system is developed nd the Plus-and Play of sensor is supported.Furthermore,we install a reformed eal-time Linux operating system on it.As a result,the existing programs on PC an be transplanted into the mentioned platform easily and there is no difference etween the development on the mentioned platform and PC.2)With the support of real-time operating system,an easy time synchronization ethod is proposed.It utilizes the PPS from GPS receiver and the local oscillator o maintain an accurate time related to standard GPS time.As a software ynchronization method,it avoids the requirement for an additional chip and its ynchronization can reach to 1 millisecond3)The error characteristics of several alignment methods are analyzed.Based on that, cascaded fast alignment strategy is proposed and it can finish the alignment rocess fast under any dynamic condition.4)Aimed at the disorder phenomena of sensor data under real-time condition,a dual-ine filter model is derived based on the sequential Kalman filter.It avoids the large mount memory spend without the loss of accuracy.In addition,the computing ime delay is processed.At the same time,the theoretical models of Zero Velocity pdate,Zero Integrated Heading Rate,Non-holonomic Constraint and Centripetal orce Constraint are derived.A factor for observation quality evaluation is given ased on the unit weight mean square error of innovation and a hierarchic information fusion strategy is proposed.5)The fusion algorithm of RTK/SINS loosely coupled system is realized and a simulation tool and a graphic user interface is developed.Three sets of real vehicle test data are used to verify the integrated navigation system.The results show that the system run properly and the RMS od position,velocity and attitude of loosely coupled mode are(0.023,0.016,0.027)m,(0.010,0.009,0.010)m/s and(1.660, 0.048,0.093)°respectively.At the same time,several GNSS signal gaps are produced to verify the auxiliary update models.The results show that the auxiliary update models can improve the navigation performance during the short time GNSS signal gaps obviously.