这是我们将autoware部署在车辆过程中写的代码注释！
https://blog.csdn.net/qq_40252459/article/details/123287437
1.配置can卡
2.可以添加激光雷达的聚类，并且画盒子表示
3.问题是current lane doesn't have change flag 节点是lane_select
Necessary topics are not subscribed yet ... 节点是pure_pursuit
4.发一个节点让can发送消息：rostopic pub -r 20 /Control_acc std_msgs/Float64 "data: 0.1"
5.128线激光雷达、将128线激光雷达加入qt界面中,将激光雷达加到autoware中，注意将launch文件改回log
6.修改qt界面中的配置文件在/home/navigation/autoware.ai/install/runtime_manager/lib/runtime_manager
7.单独编译一个节点AUTOWARE_COMPILE_WITH_CUDA=1 colcon build --cmake-args -DCMAKE_BUILD_TYPE=Release --packages-select pure_pursuit
8.如果要改寻迹点的速度，需要修改以下：
private_nh_.param("const_velocity", const_velocity_, 5.0);
9.适配速腾激光雷达：要下载rslidar_sdk，修改config中的config.yaml的frame_id为velodyne,再
10.节点waypoint_loader注释
11.最重要的发现！！！vehicle_cmd包含所有的控制消息，包含四元数，线速度，加速度，转角
12.标定命令 roslaunch autoware_camera_lidar_calibrator camera_lidar_calibration.launch intrinsics_file:=/home/ne0/Desktop/calib_heat_camera1_rear_center_fisheye.yaml compressed_stream:=True camera_id:=camera1
13.astar_avoid节点配置说明:
主要的内容在大循环while (ros::ok())中，重新规划的方法在其中有详细的描述，需要详细看，2022-0320前看完
14.relaying->stoping->planning->avoiding->relaying
15. /points_no_ground (points above ground)，是指在地面以上的点云即指障碍物，那么车是points_no_ground
16. 修改了common/op_planner/src/BehaviorStateMachine.cpp中443行的代码，使得车辆在任务结束后可以重新规划到目标点
17. 找到目标点后会输出“Current Goal Index =”
18. points_no_ground现在是有消息的。current_pose这个节点是mission_planning中某个节点发出来的。
19. /occupancy_road_status这个节点有数据了。global openplanner需要打开的节点包含：
- /initialpose和move_base_simple/goal，这是用于在rviz中输入目标点的
- /current_pose和/current_velocity是为了得到车辆现在的位姿和速度的，这个应该是在mission_planning中某个节点发出来的.
- /occupancy_road_status，得到珊格地图，这个节点打开需要勾选2个节点，在Computing下的Semantics中勾选Wayarea2grid和road_occupancy_processor，这两个节点分别发出的话题是/occupancy_wayarea和/occupancy_road_status
- 发出的话题是lane_waypoints_array，还有三个是为了在rviz中可视化global_waypoints_rviz、vector_map_center_lines_rviz和op_destinations_rviz。
- 换道逻辑在op_planner/src/MappingHelpers.cpp中的FindAdjacentLanesV2()函数中，里边作了修改，还未读懂。
- 到车辆转角公式：
a = arctan(l*w/vx) 利用公式v/r = w - ros包清理 rosclean check:检查状态 rosclean purge删除所有内容
- openplanner全局路径注释
/home/navigation/autoware.ai/src/autoware/core_planning/op_global_planner/nodes/op_global_planner_core.cpp
固定路径密度函数
/home/navigation/autoware.ai/src/autoware/core_planning/op_local_planner/nodes/op_behavior_selector/op_behavior_selector_core.cpp 同样发现了固定路径密度函数函数
1.steering_robot_有一个节点在哪？
2.问题For frame [velodyne]: No transform to fixed frame [world]. TF error: [Could not find a connection between 'world' and 'velodyne' because they are not part of the same tree.Tf has two or more unconnected trees.]
3.融合
range_vision_fusion.launch
问题：相机没有标定
4.打开ring_ground_filter和ray_ground_filter后会出现错误
[ WARN] [1640874586.661332263]: Lookup would require extrapolation into the past. Requested time 1640874461.921009664 but the earliest data is at time 1640874576.669836191, when looking up transform from frame [velodyne] to frame [base_link]
[ERROR] [1640874622.966605713]: Failed transform from base_link to velodyne
打开voxel_grid_filter后出现错误：
Failed to find match for field 'intensity'.
[ WARN] [1640876153.420807323]: Cannot get closest waypoints. All closest waypoints are changed to -1.
[ INFO] [1640876270.511634298]: Found GOAL at index = 298
[ INFO] [1640876270.517372378]: PLANNING -> AVOIDING, Found path
avoiding
[ INFO] [1640876270.310769791]: AVOIDING -> STOPPING, Abort avoiding
[ INFO] [1640876270.411066248]: STOPPING -> PLANNING, Start A* planning
planning
错误：
Error: out of memory /home/navigation/autoware.ai/src/autoware/core_perception/ndt_gpu/src/SymmetricEigenSolver.cu 13
Error: driver shutting down /home/navigation/autoware.ai/src/autoware/core_perception/ndt_gpu/src/MatrixDevice.cu 30
[ndt_mapping-2] process has died [pid 23726, exit code 1, cmd /home/navigation/autoware.ai/install/lidar_localizer/lib/lidar_localizer/ndt_mapping __name:=ndt_mapping __log:=/home/navigation/.ros/log/93ca424c-6fba-11ec-b1fd-78d00426050d/ndt_mapping-2.log].
log file: /home/navigation/.ros/log/93ca424c-6fba-11ec-b1fd-78d00426050d/ndt_mapping-2*.log
安装autoware 1.14的教程：
https://www.cnblogs.com/hgl0417/p/14617025.html
https://gitlab.com/autowarefoundation/autoware.ai/autoware/-/wikis/Source-Build
关于避障astar_avoid的网站：
https://zhuanlan.zhihu.com/p/163844578
https://blog.csdn.net/xiaoxiao123jun/article/details/111089307
关于适配速腾激光雷达：
https://blog.csdn.net/yz2630570484/article/details/114232333
关于融合：
https://blog.csdn.net/qq_40252459/article/details/124330607
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent