|
| 1 | +# LiDAR Marker Localizer |
| 2 | + |
| 3 | +**LiDARMarkerLocalizer** is a detect-reflector-based localization node . |
| 4 | + |
| 5 | +## Inputs / Outputs |
| 6 | + |
| 7 | +### `lidar_marker_localizer` node |
| 8 | + |
| 9 | +#### Input |
| 10 | + |
| 11 | +| Name | Type | Description | |
| 12 | +| :--------------------- | :---------------------------------------------- | :--------------- | |
| 13 | +| `~/input/lanelet2_map` | `autoware_map_msgs::msg::HADMapBin` | Data of lanelet2 | |
| 14 | +| `~/input/pointcloud` | `sensor_msgs::msg::PointCloud2` | PointCloud | |
| 15 | +| `~/input/ekf_pose` | `geometry_msgs::msg::PoseWithCovarianceStamped` | EKF Pose | |
| 16 | + |
| 17 | +#### Output |
| 18 | + |
| 19 | +| Name | Type | Description | |
| 20 | +| :------------------------------ | :---------------------------------------------- | :----------------------------------------------------------------- | |
| 21 | +| `~/output/pose_with_covariance` | `geometry_msgs::msg::PoseWithCovarianceStamped` | Estimated pose | |
| 22 | +| `~/debug/pose_with_covariance` | `geometry_msgs::msg::PoseWithCovarianceStamped` | [debug topic] Estimated pose | |
| 23 | +| `~/debug/marker_detected` | `geometry_msgs::msg::PoseArray` | [debug topic] Detected marker poses | |
| 24 | +| `~/debug/marker_mapped` | `visualization_msgs::msg::MarkerArray` | [debug topic] Loaded landmarks to visualize in Rviz as thin boards | |
| 25 | +| `~/debug/marker_pointcloud` | `sensor_msgs::msg::PointCloud2` | [debug topic] PointCloud of the detected marker | |
| 26 | +| `/diagnostics` | `diagnostic_msgs::msg::DiagnosticArray` | Diagnostics outputs | |
| 27 | + |
| 28 | +## Parameters |
| 29 | + |
| 30 | +{{ json_to_markdown("localization/autoware_landmark_based_localizer/autoware_lidar_marker_localizer/schema/lidar_marker_localizer.schema.json") }} |
| 31 | + |
| 32 | +## How to launch |
| 33 | + |
| 34 | +When launching Autoware, set `lidar-marker` for `pose_source`. |
| 35 | + |
| 36 | +```bash |
| 37 | +ros2 launch autoware_launch ... \ |
| 38 | + pose_source:=lidar-marker \ |
| 39 | + ... |
| 40 | +``` |
| 41 | + |
| 42 | +## Design |
| 43 | + |
| 44 | +### Flowchart |
| 45 | + |
| 46 | +```plantuml |
| 47 | +@startuml |
| 48 | +
|
| 49 | +group main process |
| 50 | + start |
| 51 | + if (Receive a map?) then (yes) |
| 52 | + else (no) |
| 53 | + stop |
| 54 | + endif |
| 55 | +
|
| 56 | + :Interpolate based on the received ego-vehicle's positions to align with sensor time; |
| 57 | +
|
| 58 | + if (Could interpolate?) then (yes) |
| 59 | + else (no) |
| 60 | + stop |
| 61 | + endif |
| 62 | +
|
| 63 | + :Detect markers (see "Detection Algorithm"); |
| 64 | +
|
| 65 | + :Calculate the distance from the ego-vehicle's positions to the nearest marker's position on the lanelet2 map; |
| 66 | +
|
| 67 | + if (Find markers?) then (yes) |
| 68 | + else (no) |
| 69 | + if (the distance is nearby?) then (yes) |
| 70 | + stop |
| 71 | + note : Error. It should have been able to detect marker |
| 72 | + else (no) |
| 73 | + stop |
| 74 | + note : Not Error. There are no markers around the ego-vehicle |
| 75 | + endif |
| 76 | + endif |
| 77 | +
|
| 78 | + :Calculate the correction amount from the ego-vehicle's position; |
| 79 | +
|
| 80 | + if (Is the found marker's position close to the one on the lanelet2 map?) then (yes) |
| 81 | + else (no) |
| 82 | + stop |
| 83 | + note : Detected something that isn't a marker |
| 84 | + endif |
| 85 | +
|
| 86 | + :Publish result; |
| 87 | +
|
| 88 | + stop |
| 89 | +end group |
| 90 | +
|
| 91 | +@enduml |
| 92 | +
|
| 93 | +``` |
| 94 | + |
| 95 | +## Detection Algorithm |
| 96 | + |
| 97 | + |
| 98 | + |
| 99 | +1. Split the LiDAR point cloud into rings along the x-axis of the base_link coordinate system at intervals of the `resolution` size. |
| 100 | +2. Find the portion of intensity that matches the `intensity_pattern`. |
| 101 | +3. Perform steps 1 and 2 for each ring, accumulate the matching indices, and detect portions where the count exceeds the `vote_threshold_for_detect_marker` as markers. |
| 102 | + |
| 103 | +## Sample Dataset |
| 104 | + |
| 105 | +- [Sample rosbag and map](https://drive.google.com/file/d/1FuGKbkWrvL_iKmtb45PO9SZl1vAaJFVG/view?usp=sharing) |
| 106 | + |
| 107 | +This dataset was acquired in National Institute for Land and Infrastructure Management, Full-scale tunnel experiment facility. |
| 108 | +The reflectors were installed by [Taisei Corporation](https://www.taisei.co.jp/english/). |
| 109 | + |
| 110 | +## Collaborators |
| 111 | + |
| 112 | +- [TIER IV](https://tier4.jp/en/) |
| 113 | +- [Taisei Corporation](https://www.taisei.co.jp/english/) |
| 114 | +- [Yuri Shimizu](https://github.com/YuriShimizu824) |
![pre-commit-ci[bot]](https://avatars.githubusercontent.com/in/68672?v=4&size=40){kind=avatar}
0 commit comments