Demos

Quick start

To get an impression of the maps wavemap can generate, you can download the maps of our RSS paper and directly view them in Rviz. To do so,

  • Choose and download one of the maps provided here

  • Open Rviz, for example by running roscore & rviz

  • Load wavemap’s rviz plugin by clicking: Add → By display type → wavemap_rviz_plugin → WavemapMap

  • In the plugin settings, under Source select File

  • Load the map you just downloaded by clicking: Loaded map → Choose file

  • [Optional] Tweak the visualization options under Render voxels and Render slice

Newer College dataset

The Newer College dataset is available here. To get the sharpest maps, we recommend supplying wavemap with a high-rate odometry estimate and turning on wavemap’s built-in pointcloud motion undistortion. In our experiments, we got these estimates by modifying FastLIO2 to publish its forward-integrated IMU poses. If you would like to run FastLIO2 yourself, our public fork is available here. Alternatively, we provide rosbags with pre-recorded odometry for the Multi-Cam Cloister, Park, Math-easy and Mine-easy sequences here.

To run wavemap on the Cloister sequence used in the paper, run:

roslaunch wavemap_ros newer_college_os0_cloister.launch rosbag_dir:=<path_to_downloaded_dataset_directory>

For additional options, please refer to the launch file’s documented arguments here. To experiment with wavemap’s configuration, modify this config file.

Panoptic mapping dataset

The Panoptic Mapping flat dataset is available here. You can automatically download it using:

export FLAT_DATA_DIR="/home/$USER/data/panoptic_mapping" # Set to path of your preference
bash <(curl -s https://raw.githubusercontent.com/ethz-asl/panoptic_mapping/3926396d92f6e3255748ced61f5519c9b102570f/panoptic_mapping_utils/scripts/download_flat_dataset.sh)

To process it with wavemap, run:

roslaunch wavemap_ros panoptic_mapping_rgbd_flat.launch base_path:="${FLAT_DATA_DIR}"/flat_dataset/run1

To experiment with different wavemap settings, modify this config file.

Your own data

The only requirements for running wavemap are:

  1. a source of depth measurements,

  2. sensor pose (estimates) for each measurement.

We usually use depth measurements from depth cameras or 3D LiDARs, but any source would work as long as a sufficiently good projection and measurement model is available. Wavemap’s ROS interface natively supports depth image and pointcloud inputs, and automatically queries the sensor poses from the TF tree.

To help you get started quickly, we provide various example config and ROS launch files. For a brief introduction on how to edit wavemap configs and an overview of all the available settings, see the configuration page.

Multi-sensor live demo

This section provides instructions to reproduce the interactive multi-sensor, multi-resolution mapping demo we performed at several events, including RSS 2023 and the Swiss Robotics Day 2023. In this demo, wavemap fuses measurements from a depth camera up to a resolution of 1cm and a LiDAR up to a range of 15m in real-time on a laptop. The odometry is obtained by running FastLIO2 using the LiDAR’s pointclouds and built-in IMU.

Hardware

Wavemap can be configured to work with almost any depth input. In case you want to replicate our exact setup, we used a Livox MID-360 LiDAR at RSS and an Ouster OS0-128 at the Swiss Robotics Day. We chose the PMD Pico Monstar depth camera for both demos, but the Azure Kinect also works well. Although any recent laptop will suffice, we recommend using a laptop with a discrete GPU as it will help Rviz run much smoother.

Docker

For convenience, the entire demo can be run in Docker using this Docker image, which includes the sensor drivers, FastLIO and wavemap. You can build it by running:

docker build --tag=wavemap_demo --pull - <<< $(curl -s https://raw.githubusercontent.com/ethz-asl/wavemap/main/tooling/docker/live_demo.Dockerfile)

We provide a convenience script that you can use to run the Docker image. To download it, run:

curl -o demo_in_docker.sh https://raw.githubusercontent.com/ethz-asl/wavemap/main/tooling/packages/wavemap_utils/scripts/demo_in_docker.sh
sudo chmod +x demo_in_docker.sh

Configure

Start by configuring the laptop’s network interface to receive the LiDAR data. By default, the Livox we used required the laptop’s IP to be set to 192.168.1.50, while the Ouster required the laptop’s IP to be set to 192.168.10.50. In case you’re using another LiDAR, the expected receiver IP is usually mentioned in the LiDARs manual or driver documentation.

Next, update the sensor calibrations to match your sensor setup. If you’re using the exact sensors we used, you mainly need to update the extrinsics. These are defined through static TFs, at the bottom of the launch file for your sensor setup (e.g. here or here). If you’re using different sensors, you will also need to update the wavemap, FastLIO and sensor driver config files. To help you get started quickly, we provide examples for many different sensor setups for wavemap and FastLIO. In case we do not yet have an example FastLIO config for your LiDAR, you might find one in the official the FastLIO repository.

Note that the demo_in_docker.sh script saves and restores the ros/wavemap_ros folder across runs (explained here), so you can configure wavemap by simply running demo_in_docker.sh and editing the launch and config files through the resulting interactive shell.

Run

Once the configured, you can directly run the demo by calling the demo_in_docker.sh script followed by the launch command. For example:

demo_in_docker.sh roslaunch wavemap_ros ouster_os0_pico_monstar.launch