Skip to content

Commit

Permalink
update to v2.0.1
Browse files Browse the repository at this point in the history
  • Loading branch information
lingbomeng committed Nov 19, 2024
1 parent 610554a commit a0ba7b7
Show file tree
Hide file tree
Showing 7 changed files with 109 additions and 175 deletions.
3 changes: 2 additions & 1 deletion .gitignore
Original file line number Diff line number Diff line change
@@ -1 +1,2 @@
.vscode
.vscode
./PCD/scans.pcd
1 change: 1 addition & 0 deletions PCD/do_not_delete_this_file.txt
Original file line number Diff line number Diff line change
@@ -0,0 +1 @@
This file is used to maintain the existence of this directory. Otherwise, the PCD file can not be saved!
Binary file added PCD/scans.pcd
Binary file not shown.
54 changes: 46 additions & 8 deletions README.md
Original file line number Diff line number Diff line change
Expand Up @@ -2,9 +2,11 @@

## 1. Introduction

### 1.1 Unitree LiDAR

This repository adapts the state-of-the-art lidar inertial odometry algorithm, `Point-LIO`, for use with our lidar products:
- `Unitree Lidar L1`
- `Unitree Lidar L2`
- `Unitree LiDAR L1`
- `Unitree LiDAR L2`

Both `L1` and `L2` possess these features:
- large field of view (360° × 90°)
Expand All @@ -13,20 +15,24 @@ Both `L1` and `L2` possess these features:
- suitable for applications in low-speed mobile robots

If you want to learn more about our lidar products, you can refer to the official website for details.
- <https://m.unitree.com/en/LiDAR/>
- <https://www.unitree.com/L2>
- <https://www.unitree.com/LiDAR>


### 1.2 Point-LIO

`Point-LIO` is a robust and high-bandwidth lidar inertial odometry (LIO) with the capability to provide accurate, high-frequency odometry and reliable mapping under severe vibrations and aggressive motions. If you need further information about the `Point-LIO` algorithm, you can refer to their official website and paper:
- <https://github.com/hku-mars/Point-LIO>
- [Point‐LIO: Robust High‐Bandwidth Light Detection and Ranging Inertial Odometry](https://onlinelibrary.wiley.com/doi/epdf/10.1002/aisy.202200459)


## 2. Video Demos

### 2.1 L1 Lidar
### 2.1 L1 LiDAR

[![Video](./doc/video.png)](https://oss-global-cdn.unitree.com/static/c0bd0ac7d1e147e7a7eaf909f1fc214f.mp4 "SLAM based on Unitree 4D LiDAR L1")

### 2.2 L2 Lidar
### 2.2 L2 LiDAR

[![Video](./doc/l2-demo-video-bilibili.png)](https://www.bilibili.com/video/BV1XVUVYHEHR "SLAM based on Unitree 4D LiDAR L2")

Expand Down Expand Up @@ -95,7 +101,7 @@ catkin_make

## 5. Run

### 5.1 Run with Unilidar L1
### 5.1 Run with L1

To ensure proper initialization of the IMU, it is advisable to keep the lidar in a stationary state during the initial few seconds of algorithm execution.

Expand Down Expand Up @@ -128,7 +134,7 @@ You can use the `pcl_viewer` tool to view this pcd file:
pcl_viewer scans.pcd
```

### 5.2 Run with Unilidar L1 dataset
### 5.2 Run with rosbag of L1

If you don't have our lidar for now, you can download our dataset recorded with our lidar and run testify this algorithm with it.
The download address is here:
Expand Down Expand Up @@ -160,7 +166,7 @@ You can use the `pcl_viewer` tool to view this pcd file:
pcl_viewer scans.pcd
```

### 5.3 Run with Unilidar L2
### 5.3 Run with L2

To ensure proper initialization of the IMU, it is advisable to keep the lidar in a stationary state during the initial few seconds of algorithm execution.

Expand All @@ -187,6 +193,38 @@ After completion of the run, all cached pointcloud map will be saved to the foll
catkin_point_lio_unilidar/src/point_lio_unilidar/PCD/scans.pcd
```

You can use the `pcl_viewer` tool to view this pcd file:
```
pcl_viewer scans.pcd
```

### 5.4 Run with rosbag of L2

If you don't have our lidar for now, you can download our dataset recorded with our lidar and run testify this algorithm with it.
The download address is here:
- [L2 Indoor Point Cloud Data.bag - Download](https://oss-global-cdn.unitree.com/static/L2%20Indoor%20Point%20Cloud%20Data.bag)
- [L2 Park Observed Point Cloud Data.bag - Download](https://oss-global-cdn.unitree.com/static/L2%20Park%20Point%20Cloud%20Data.bag)


Run `Point-LIO`:
```
cd catkin_point_lio_unilidar
source devel/setup.bash
roslaunch point_lio_unilidar mapping_unilidar_l2.launch
```

Play the dataset you downloaded:
```
rosbag play XXXXXX.bag
```

After completion of the run, all cached pointcloud map will be saved to the following path:
```
catkin_point_lio_unilidar/src/point_lio_unilidarPCD/scans.pcd
```

You can use the `pcl_viewer` tool to view this pcd file:
```
pcl_viewer scans.pcd
Expand Down
6 changes: 6 additions & 0 deletions VERSION.md
Original file line number Diff line number Diff line change
Expand Up @@ -12,3 +12,9 @@

## v2.0.0 (2024.11.14)
- Add support for Unitree Lidar L2

## v2.0.1 (2024.11.19)
- Add directory `PCD` for saving pcd files
- Add odometry topic
- Modify topic names to unified routines.
- Add download address for L2 rosbag files
Loading

0 comments on commit a0ba7b7

Please sign in to comment.