Skip to content

L3+ autonomous driving perception pre-research

10/2020: L3+pre-research, Kick-off meeting of the L3+ autonomous driving perception pre-research project based on DDS communication, with interns Ruichen Tan (Ohio University), Hao Chao, Zhigao He,Peijie Chen, etc. Four major categories of things were accomplished:

  • Established a communication matrix and completed the hierarchical decomposition of software modules.
  • L3+ implements CAN DBC parsing and successfully integrates the L3- fused simulink model (c code generated using a code generation tool).
  • Completed the obstacle detection of visual YOLO and lidar YOLO
  • Draw the laser SLAM-NDT to obtain the point cloud map of the Dongfeng technology center, and use RoadRunner to draw the map to obtain the road network OpenDrive map. Fusion loaction through IMU+GPS.

开发成员：严义雄，曾超豪/何高志/陈培杰/黄宇航（联友科技），谭睿宸（实习）

从信号链路上打通了“视觉，激光，毫米波，IMU，GPU和地图，7大类信号

开发目的：

  • 【通信矩阵建立】实现感知框架（架构+底层驱动+IO）

  • 【L3- Node】实现加载L2/L3- model并进行online/offline测试）

  • 具体为：支持替换DBC

  • 具体为：支持替换L3 Node实现上述要求

  • 【障碍物检测】实现面向L3+的视觉&激光雷达融合

  • 具体为：yolo前视摄像头，yolo激光雷达

  • 具体为：车体box检测

  • 【OpenDrive格式地图】实现RTK+Lidar高精度建图

  • 具体为：OpenRoadEd实现激光点云到路网的构建
  • OpenDrive的Node节点，IMU+GPS融合节点

自动驾驶中的任务

image-20240314165456524

image-20240314165530998

演示动画：

预研L3+

逻辑框图

image-20240229101441718

OpenDrive园区原型图

img

技术中心点云地图：

image-20240229100721986