CN-122021079-A - High-speed multi-scene vehicle-road collaborative driving simulation system

CN122021079ACN 122021079 ACN122021079 ACN 122021079ACN-122021079-A

Abstract

The application relates to the technical field of road condition simulation, in particular to a high-speed multi-scene vehicle road collaborative driving simulation system, which comprises a scene simulation and data acquisition module, a simulation module and a control module, wherein the scene simulation and data acquisition module acquires point cloud data around a simulated vehicle in real time through a virtual simulation scene; the system comprises a noise analysis module, a point cloud filtering module, a vehicle-road cooperation module and a vehicle-road cooperation module, wherein the noise analysis module is used for analyzing local and global outlier characteristics of all point cloud data in each voxel and calculating noise interference prominence of each point cloud data, the point cloud filtering module is used for identifying and eliminating abnormal point clouds based on the noise interference prominence, voxel filtering processing is carried out on the rest point cloud data, and the filtered point cloud data is used for updating a three-dimensional point cloud map of a simulated vehicle to realize vehicle-road cooperation simulation test. The method aims to improve the effect of the voxel filtering algorithm on filtering processing of the point cloud data, so that the quality of updating the three-dimensional point cloud map in real time is ensured.

Inventors

GUO ZHIJIE
Wang Enshi
YANG TAO
DENG MIN
YANG GE
XIAO YIKUN

Assignees

武汉中交交通工程有限责任公司

Dates

Publication Date: 20260512
Application Date: 20260414

Claims (10)

1. A high-speed multi-scene vehicle-road co-driving simulation system, the system comprising: the scene simulation and data acquisition module is used for acquiring point cloud data around the simulated vehicle in real time through a virtual simulation scene; The noise analysis module is used for carrying out voxel grid division on point cloud data, analyzing local and global outlier characteristics of all the point cloud data in each voxel, forming two-dimensional characteristic points, acquiring a neighborhood point cloud set of each point cloud data in each voxel by utilizing the difference distance between the two-dimensional characteristic points, and calculating the noise interference prominence of each point cloud data by combining the difference distance between the two-dimensional characteristic points and the two-dimensional characteristic origin; The point cloud filtering module is used for calculating the interfered credibility of each point cloud data based on the noise interference prominence and the proportion of the noise interference prominence exceeding the segmentation threshold value part so as to identify and reject abnormal point clouds; and the vehicle-road cooperative module is used for updating the three-dimensional point cloud map of the simulated vehicle by using the filtered point cloud data so as to realize the vehicle-road cooperative simulation test.
2. The high-speed multi-scene vehicle-road collaborative driving simulation system according to claim 1, wherein the virtual simulation scene is built by PreScan simulation software.
3. A high-speed multi-scene vehicle co-driving simulation system according to claim 1, wherein the local outlier feature calculates the local outlier factor of each point cloud data in each voxel by using LOF algorithm.
4. A high speed multi-scene vehicle co-driving simulation system according to claim 1, wherein the global outlier is determined by calculating an average of absolute differences of local outlier factors of each point cloud data and all point cloud data.
5. The high-speed multi-scene vehicle-road collaborative driving simulation system according to claim 1, wherein the two-dimensional feature points consist of local outlier features and global outlier features after range normalization.
6. The high-speed multi-scene vehicle-road collaborative driving simulation system according to claim 1, wherein the obtaining the neighborhood point cloud set of each point cloud data in each voxel comprises taking two-dimensional feature points of all point cloud data in all voxels as input of an OPTICS clustering algorithm, and outputting the neighborhood point cloud set of each point cloud data in each voxel.
7. The high-speed multi-scene vehicle-road collaborative driving simulation system according to claim 1, wherein the noise interference prominence is positively correlated with a difference distance between a two-dimensional feature point and a two-dimensional feature origin of each point cloud data and is negatively correlated with the number of point clouds in a neighborhood point cloud set of each point cloud data.
8. The high-speed multi-scene vehicle-road collaborative driving simulation system according to claim 1, wherein the segmentation threshold is obtained by adaptively calculating noise interference prominence of all point cloud data in a current point cloud space through a maximum inter-class variance algorithm.
9. The high-speed multi-scene vehicle-road collaborative driving simulation system according to claim 1, wherein the method for identifying and eliminating abnormal point clouds is characterized in that the interfered credibility of all point cloud data is taken as input, the upper quartile is counted, and the point cloud data higher than the upper quartile is marked as the abnormal point cloud and eliminated.
10. The high-speed multi-scene vehicle-road collaborative driving simulation system according to claim 1, wherein during the voxel filtering processing, all the remaining point cloud data in each voxel are used as input of a voxel filtering algorithm, the centroid of all the remaining point cloud data in each voxel is calculated as a representative point of each voxel through the voxel filtering algorithm, and the voxel filtering algorithm outputs a point cloud data set after the filtering processing in the current point cloud space.

Description

High-speed multi-scene vehicle-road collaborative driving simulation system Technical Field The application relates to the technical field of road condition simulation, in particular to a high-speed multi-scene vehicle road collaborative driving simulation system. Background The vehicle-road cooperation technology (cooperative vehicle-infra structure system, CVIS) is used as an important technology for intelligent traffic development, adopts technologies such as wireless communication, new generation Internet and the like, carries out dynamic information interaction among vehicles, vehicles and roads and between vehicles and people in all directions, fully realizes effective cooperation of people, vehicles and roads, provides a reliable implementation way for improving traffic efficiency and enhancing traffic safety, and enables traffic flow to run efficiently, orderly and safely. At present, a vehicle-road collaborative driving simulation system is built based on a six-degree-of-freedom motion platform, three-degree-of-freedom translational motion and three-degree-of-freedom rotational motion simulation kinetic energy can be provided, vehicle-road collaborative driving simulation tests of various traffic scenes in a highway can be developed, and therefore the traffic safety and the traffic fluency of vehicle-road collaborative driving are verified, and the collaborative efficiency between vehicles and road measures is further improved. In the prior art, a highway simulation environment and a virtual sensor are constructed through simulation software, point cloud data around a simulated vehicle is acquired in real time by using the virtual sensor in vehicle-road collaborative driving simulation, and discrete coordinates of the point cloud data under a map coordinate system are acquired in a coordinate system conversion mode, so that a three-dimensional point cloud map of the simulated vehicle is updated in real time, and the traffic safety and the traffic fluency of vehicle-road collaborative driving are facilitated to be improved. However, since the point cloud data acquired in real time contains redundant point cloud data or noise point cloud data, in the prior art, a voxel filtering algorithm is generally adopted to carry out filtering processing on the point cloud data acquired in real time, but in the traditional voxel filtering algorithm, the centroid of all the point cloud data in a voxel is generally directly calculated as a representative point of the voxel, and abnormal point clouds generated by the influence of noise in the voxel can influence the representative point selection of the voxel, the abnormal point clouds generated by the influence of noise in the voxel are not accurately identified and removed in the prior art, so that the accuracy of voxel representative point selection is poor, the quality of a three-dimensional point cloud map is further influenced, and the traffic safety and the traffic fluency of vehicle-road collaborative driving are not facilitated. Disclosure of Invention In order to solve the technical problems, the application aims to provide a high-speed multi-scene vehicle-road collaborative driving simulation system, which adopts the following technical scheme: the application provides a high-speed multi-scene vehicle-road collaborative driving simulation system, which comprises: the scene simulation and data acquisition module is used for acquiring point cloud data around the simulated vehicle in real time through a virtual simulation scene; The noise analysis module is used for carrying out voxel grid division on point cloud data, analyzing local and global outlier characteristics of all the point cloud data in each voxel, forming two-dimensional characteristic points, acquiring a neighborhood point cloud set of each point cloud data in each voxel by utilizing the difference distance between the two-dimensional characteristic points, and calculating the noise interference prominence of each point cloud data by combining the difference distance between the two-dimensional characteristic points and the two-dimensional characteristic origin; The point cloud filtering module is used for calculating the interfered credibility of each point cloud data based on the noise interference prominence and the proportion of the noise interference prominence exceeding the segmentation threshold value part so as to identify and reject abnormal point clouds; and the vehicle-road cooperative module is used for updating the three-dimensional point cloud map of the simulated vehicle by using the filtered point cloud data so as to realize the vehicle-road cooperative simulation test. Preferably, the virtual simulation scene is built through PreScan simulation software. Preferably, the local outlier feature calculates the local outlier factor for each point cloud data within each voxel by using a LOF algorithm. Preferably, the global outlier feature is determined by calculating an average of