Research On Vehicle Lidar Modeling And Application

With the development of the city,the road traffic environment is getting worse.Vehicle intelligence has become an inevitable development trend.As an indispensable assistance in the process of intelligence,various sensors’ selection and application are also crucial.Lidar can return a large number of data by transmitting and receiving the laser beam.It has a high angle resolution and ranging accuracy.It can achieve the needs of target detection,recognition,classification and tracking in intelligent vehicles.It can be said that lidar is a powerful assistance for vehicle intelligent industry.With the increasing attention to the safety requirements of the intelligent functions of vehicle,the development and verification of the functions become more and more difficult.The solution to this problem is to build a virtual model.In simulation environment,the virtual model of lidar is used to develop intelligent vehicles.On the other hand,from the perspective of lidar application,it is expensive and greatly affected by the environment,so we can study its detection mechanism through the lidar model and analyze the factors that affect lidar detection and imaging,which is more convenient and efficient.In this paper,the generation method of lidar point cloud and factors affecting its generation are studied.Then,the vehicle intelligence is studied by using its model.Firstly,according to the detection mechanism of lidar,analyze the factors affecting the detection,specifically analyze the laser attenuation equation and the causes of physical noise.Simulate the generation of point cloud according to the detection mechanism of lidar,the physical shielding and reflection properties of laser.Combination with the factors of shadow lidar detection,establish the physical attenuation model and raindrop noise that affect the generation of point cloud Model.The main research contents are as follows:1)Analysis of lidar detection mechanism and influencing factors.Starting from the structure and detection mechanism of lidar,this paper analyzes the factors of attenuation and noise that affect its detection,and analyzes the detection power equation of lidar.Combining with Mie scattering and other theories,the general expression of lidar detection range performance and attenuation coefficient is obtained and the formation of raindrop noise is decomposed into probability expression,which lay foundations for the establishment of physical attenuation model and raindrop noise model.2)The establishment of ideal point cloud model and physical attenuation model of lidar.The representation of the model is in the form of point cloud,so the representation of point cloud should be completed firstly: establish a simple target geometry model and lidar scanning area model.Obtain the coordinate representation of the target model relative to lidar through coordinate system transformation Judge whether the surface of the target geometry model is visible to lidar,and generate the model on the visible surface of the target by combining the laser beam equation According to the analysis of attenuation,combined with the specific parameters and particle distribution model in different weather,the attenuation model of lidar is obtained.Finally,the simulation verification of point cloud generation and the data verification of attenuation model.3)The establishment of raindrop noise model of lidar.In rainy days,the point cloud detected by lidar will contain the nontarget point cloud,which is the raindrop noise generated by raindrops detected by lidar.According to the analysis of the generation mechanism of the noise,the establishment conditions of the noise are obtained.The Monte Carlo method and the rejection sampling method are used to combine with the scale distribution simulation to complete the data sampling.Then the probability of the noise is calculated based on the sampling data of the lidar in rainy days.Further,it is verified by the noise in heavy rain,and then the position distribution of the noise is calculated.Finally,the simulation of the noise is carried out.4)Application of lidar model in vehicle intelligent driving.Using the clustering and tracking algorithm of point cloud to extract the target information,and then integrate it with the lidar model into the simulation software to provide various target vehicle parameter information for the vehicle.Verify the point cloud generated by lidar model in different weather conditions.At the same time,the control algorithm of the vehicle adaptive cruise system is designed to complete the simulation of the design scene under different weather conditions,which verifies the application of lidar in vehicle intelligent driving and the feasibility of the related control algorithm.