Research On Intelligent Vehicle Navigation Algorithm Based On Multi-sensor Fusion

With the rise of the new generation of artificial intelligence technology,technologies such as deep learning,big data,cloud computing are increasingly applied in various control systems,among which unmanned driving has become a research hotspot at home and abroad in recent years.The technologies involved in unmanned driving are relatively complex,among which the core problem is the design of navigation algorithm.There are lidar,visual slam,deep learning,image recognition and so on in the navigation algorithm.In order to simplify the research on the navigation problem in unmanned driving,this paper takes intelligent car as the research platform,and proposes the intelligent vehicle navigation algorithm based on multi-sensor fusion on the basis of the detailed analysis and research of various sensing sensors and algorithms.The algorithm collects ultrasonic,infrared,GPS,lidar and other sensor information synthetical,and carries on the fusion calculation to the above information,and finally completes the path planning.Because the data of lidar sensor is relatively complex,the paper focuses on analyzing and studying the data processing of the sensor,and studies the core algorithm of scan matching of low-cost lidar.In the path planning design,A * and D * algorithms are integrated and improved.The working principles of different laser radar sensor modules are analyzed in detail,the theoretical basis of selecting slamtec mapper sensor module is demonstrated,and the rough search fusion navigation model of intelligent vehicle in campus environment is proposed combining slamtec mapper sensor data and GPS data.Finally,by analyzing the principle of the relevant navigation algorithm and the external sensor module in detail,the design of the overall structure of the intelligent car is demonstrated and put forward.