Research On Intelligent Vehicle Navigation And Location In Urban Complex Environment

The future intelligent vehicle navigation system should become accurate, steady and reasonably priced. To achieve this goal, many fusion models have been built(GNSS/DR, GNSS/INS, GNSS/MM). Although some have been proved successfully used in different kinds of environment, they still have shortages, especially in areas where the accuracy of GNSS is jeopardized. The method of researching intelligent vehicle navigation in complex environment has become an attractive research field. This paper systematically studied the variety of complex urban environment influencing the result of navigation and position. Then take the corresponding navigation method in different environment. Finally, we verify the correctness and feasibility of proposed method through real data which being collected from experiment.The main contents of this thesis are divided into six parts:1. Dividing the complex urban environment into the fuzzy road condition and the urban canyon condition, and then build different navigating and positioning method as well as models to improve the localization accuracy under the different complex environment.2. Present an improved map matching algorithm based on the weights of WΔω, Wd and Wθ, and Choose the GPS points’ best matching road among the candidates through integrating of these weights. Finally, after comparing the raw trajectory with corrected trajectory by software simulation based on MATLAB and Visual C++, this new algorithm is proved to be appropriate for the ambiguous environment which has mentioned before. The simulation results reflect the correcting of the navigation trajectory error.3. Studied the theory of traditional Stereo-Vision measurement. Utilize Zhang’s calibration method to calibrate camera’s internal parameters and external parameters. A localization algorithm has been designed under the visual coordinate system. In the end, analyze the error of the visual measurements.4. Designing a Stereo-Vision assisted GNSS localization method which based on landmarks. An assisted position process system has been built, including landmark detecting, visual calculate and assisted position. Random Hough Transform(RHT) has been used to detect the landmark, SIFT and K-means algorithm use to recognize and match landmarks. Stereo-disparity figures out the vector between vehicle and landmark, then build an aided model to calculate vehicle’s position.5. Designing the software for cameras to collect the images, the contents include using experimental vehicle collect the Real-time data in a challenged environment. Finally, after analyzing the error of Stereo-Vision and the error of GNSS, we prove the proposed method make an obvious improvement on GNSS localization results in the areas where landmarks are visible.6. Research the Kalman filter algorithm and integrate the visual information with GNSS information to build a novel Vision/GNSS integrated navigation data fusing model.