Research On Intelligent Vehicle Navigation And Position Algorithms Based On GPS/SINS/Odometer

Intelligent vehicle has become a research hotspot at home and abroad,it has multiple functions such as the environmental perception,decision-making and planning,auxiliary driving.Based on the intelligent vehicle of Beijing University of Technology,the dissertation focuses on the intelligent vehicle navigation and position technology and filter algorithms,the main contents are as follows:(1)Aiming at eliminating the initial error of strap-down inertial navigation system(SINS),and to provide the inertial navigation calculation with initial attitude matrix and attitude angles,the initial alignment on stationary base based on reduced order Kalman filter is seriously researched and analyzed.The two-wheeled cart with inertial measurement unit is used to verify the initial alignment algorithm,the results show that the initial alignment based on reduced order Kalman filter algorithm can accurately estimate the small misalignment angles and obtain the accurate initial attitude matrix and attitude angles.(2)To guarantee the precision of inertial navigation calculation and eliminate the influence of high frequency noise and vibration,two de-noising methods,respectively based on wavelet transform and adaptive fading Kalman filter algorithm based on current statistical model,are proposed to preprocess the dynamic data of SINS.Then the dissertation uses the original and de-noised data to carry on the inertial navigation calculation.The position results prove that after de-noising,the position error is greatly reduced,and the adaptive fading Kalman filter algorithm based on current statistical model is with higher position accuracy and easy to be realized.(3)During the GPS outage for a long time,influenced by the precision of inertial sensors and integral working principle of SINS,only using the pure inertial navigation and positioning method can lead to positioning error accumulation,so the SINS/odometer integrated navigation scheme based on the fuzzy adaptive Kalman filter after the filter abnormal judgment is proposed to adjust the measurement noise covariance matrix and implement the optimal estimation of the status value.When the GPS signal is available,according to the essence of the nonlinear system,the GPS/SINS nonlinear integrated navigation scheme based on the simplified unscented Kalman filter is seriously studied,the nonlinear system model based on large initial misalignment angles is established,the both process noise and measurement noise were complex additive noise with linear measurement equation.And then,the simplified unscented Kalman filter is used to realize the nonlinear state error estimate for GPS/SINS integrated navigation.(4)The navigation sensors equipped on the intelligent vehicle and their key parameters relative to programming are introduced.Also to guarantee the time synchronization of data,which are used to implement integrated navigation,we design the integrated navigation and positioning software platform.With the designed software,we collect the dynamic data of sensors in real vehicle experiment and complete the validation of the two integrated navigation schemes: SINS/odometer integrated navigation and GPS/SINS nonlinear integrated navigation,the position results show that the proposed integrated navigation schemes are with higher position accuracy,much more simple and with smaller amount of calculation,which can satisfy the requirement of intelligent vehicle navigation and position.