ROS Based Intelligent Vehicle Autonomous Navigation System Design

With the rapid development of modern science and technology,high-tech has been widely used in our life,which makes people's daily life become more efficient and convenient.In recent years,there appeared high-tech products such as fire drones,unmanned speedboats,intelligent robots,intelligent unmanned vehicles.Particularly speaking,the emergence of intelligent unmanned vehicles has lowered the threshold for driving cars.With more and more people studying unmanned vehicles,autonomous navigation system of intelligent vehicles has become a hot topic in recent years.Although there have been many scientific achievements in the intelligent navigation of unmanned vehicles in known maps,the positioning and navigation of unmanned vehicles in the case of unknown environmental maps are still the main research interests.The Simultaneous Localization And Mapping(SLAM)algorithm can measure the information of the unfamiliar environment and estimate its position through the sensors carried by the intelligent vehicle itself,which can provide effective positioning and navigation for the intelligent vehicle in the unfamiliar environment.The precise positioning and navigation of intelligent vehicles in unfamiliar environments is a prerequisite for the popularity of intelligent unmanned vehicles.Therefore,in-depth study of SLAM algorithm is of great research significance and commercial value for improving the positioning and navigation technology of intelligent unmanned vehicles.The current intelligent unmanned vehicles mainly have the following problems in unfamiliar environments:their own positioning accuracy is poor;the global path planning effect is not ideal;there is no autonomous navigation control strategy.In order to solve the above-mentioned problems of smart cars,this paper starts from the hardware and software aspects to design an autonomous navigation system for smart unmanned vehicles that can meet positioning and navigation in real life.main tasks as follows:(1)In terms of the hardware,with the help of the laser radar,millimeter wave radar,vision sensors and inertial navigation equipped on intelligent unmanned vehicles,the information of unfamiliar environment of the intelligent unmanned vehicles is collected,which is conducive to draw the high-precision environment map that can meet the actual use.Together with the improved SLAM algorithm described below,the positioning accuracy of the intelligent unmanned vehicles can be improved.(2)In terms of the positioning algorithm,firstly,for the disadvantages of the traditional extended Kalman filter SLAM(EKF-SLAM)algorithm with low universality and low accuracy,an extended Kalman filter structure based on fuzzy adaptive control is given.SLAM algorithm,through the introduction of fuzzy adaptive control module in the EKF-SLAM process,the accuracy of the intelligent vehicle's pose estimation can be effectively improved.Secondly,the traditional FastSLAM algorithm is based on particle filter estimation,so the error between the estimated pose and the real pose will increase with the movement.Therefore,an improved algorithm combined with historical particles is proposed to compensate for the pose error caused by the movement.(3)In terms of navigation algorithm,the autonomous navigation function of the intelligent vehicle is realized by a combination of the A*algorithm of global path planning and DWA algorithm of the local path planning.(4)In this paper,two common autonomous navigation control strategies are designed.The strategy of waypoint storage+ACC+AEB+single point preview+positioning is used in the ROS system to realize the autonomous return function of the intelligent vehicle.The strategy of SLAM+planning+ACC+AEB+single point preview+Positioning is used to realize automatic driving function.Finally,the effectiveness of autonomous navigation system of the intelligent vehicle is verified by the field experiment on the intelligent vehicle platform.The experimental results show that two strategies and the improved algorithms can fully meet the needs of unmanned driving on campus.