Home

Awesome

UrbanLoco

UrbanLoco: A Full Sensor Suite Dataset for Mapping and Localization in Urban Scenes

Abstract

Mapping and localization is a critical module of autonomous driving, and significant achievements have been reached in this field. Beyond Global Navigation Satellite System (GNSS), research in point cloud registration, visual feature matching, and inertia navigation has greatly enhanced the accuracy and robustness of mapping and localization in different scenarios. However, highly urbanized scenes are still challenging: LIDAR- and camera-based methods perform poorly with numerous dynamic objects; the GNSS-based solutions experience signal loss and multipath problems; the inertia measurement units (IMU) suffer from drifting. Unfortunately, current public datasets either do not adequately address this urban challenge or do not provide enough sensor information related to mapping and localization. Here we present UrbanLoco: a mapping/localization dataset collected in highly-urbanized environments with a full sensor-suite. The dataset includes 13 trajectories collected in San Francisco and Hong Kong, covering a total length of over 40 kilometers. Our dataset includes a wide variety of urban terrains: urban canyons, bridges, tunnels, sharp turns, etc. More importantly, our dataset includes information from LIDAR, cameras, IMU, and GNSS receivers.

Keywords: Mpapping, Localization, Urban Areas, Full Sensor Suit, Hong Kong, San Francisco

<p align="center"> <img width="pix" src="img/ppp1.png"> </p>

Important Notes:

Contact Authors:

Related Papers :

if you use UrbanLoco for your academic research, please cite our paper.

Work related to urbanLoco Dataset :

1. California Dataset

1.1 Sensor Setups

The platform for data collection in Hong Kong is a Honda Fit. The platform is equipped with the following sensors:

<p align="center"> <img width="712pix" src="img/prius.png"> </p>

Intrinsic and Extrinsic Parameters, ROSBAG Information

The coordinates transformation between multiple sensors, and intrinsic parameters of camera can be found via Intrinsic and Extrinsic Parameters.

1.2. Dataset 1: CABayBridge20190828151211

Brief: Dataset CABayBridge20190828151211 is collected near Bay Bridge of San Francisco.

Some key features are as follows:

Date of CollectionTotal SizeSensorsDownloadFeatures
2019/08/2844.6 GBGNSS/LiDAR/Camera/IMU/SPAN-CPTGoogleDriveDynamic Objects, Sharp Turn
<!-- | | | | | | | **GNSS** | **IMU** |**LiDAR** |**Camera** | **Ground Truth**| | ```/ublox_gps_node/fix``` | ```/imu_raw``` |```/rslidar_points``` |```/camera_array/cam0/image_raw/compressed``` | ``/novatel_data/bestpos``| --> <!-- **Rostopic for camera**: we provide camera images from six cameras with rostopic name from ```/camera_array/cam0/image_raw/compressed``` to ```/camera_array/cam5/image_raw/compressed``` --> <p align="center"> <img width="420pix" src="img/CABayBridge20190828151211.png"> </p>

1.3. Dataset 2: CAMarketStreet20190828155828

Brief: Dataset CAMarketStreet20190828155828 is collected near market street of San Francisco.

Some key features are as follows:

Date of CollectionTotal SizeSensorsDownloadFeatures
2019/08/2860.6 GBGNSS/LiDAR/Camera/IMU/SPAN-CPTGoogleDriveDynamic Objects, high-rising buildings
<p align="center"> <img width="420pix" src="img/CAMarketStreet20190828155828.png"> </p>

1.4. Dataset 3: CARussianHill20190828173350

Brief: Dataset CARussianHill20190828173350 is collected near Bay Bridge of San Francisco.

Some key features are as follows:

Date of CollectionTotal SizeSensorsDownloadFeatures
2019/08/2867.2 GBGNSS/LiDAR/Camera/IMU/SPAN-CPTGoogleDriveDynamic Objects, high-rising buildings
<p align="center"> <img width="420pix" src="img/CARussianHill20190828173350.png"> </p>

1.5. Dataset 4: CAChinaTown20190828180248

Brief: Dataset CAChinaTown20190828180248 is collected near a China Town of San Francisco.

Some key features are as follows:

Date of CollectionTotal SizeSensorsDownloadFeatures
2019/08/2854.3 GBGNSS/LiDAR/Camera/IMU/SPAN-CPTGoogleDriveDynamic Objects, high-rising buildings
<p align="center"> <img width="420pix" src="img/CAChinaTown20190828180248.png"> </p>

1.6. Dataset 5: CAColiTower20190828184706

Brief: Dataset CAColiTower20190828184706 is collected near Coli Tower of San Francisco.

Some key features are as follows:

Date of CollectionTotal SizeSensorsDownloadFeatures
2019/08/289.73 GBGNSS/LiDAR/Camera/IMU/SPAN-CPTGoogleDriveDynamic Objects, high-rising buildings
<p align="center"> <img width="320pix" src="img/CAColiTower20190828184706.png"> </p>

1.7. Dataset 6: CALombardStreet20190828190411

Brief: Dataset CALombardStreet20190828190411 is collected near Lombard street of San Francisco.

Some key features are as follows:

Date of CollectionTotal SizeSensorsDownloadFeatures
2019/08/289.83 GBGNSS/LiDAR/Camera/IMU/SPAN-CPTGoogleDriveDynamic Objects, high-rising buildings
<p align="center"> <img width="420pix" src="img/CALombardStreet20190828190411.png"> </p>

1.8. Dataset 7: CAGoldenBridge20190828191451

Brief: Dataset CAGoldenBridge20190828191451 is collected near Golden Bridge of San Francisco.

Some key features are as follows:

Date of CollectionTotal SizeSensorsDownloadFeatures
2019/08/2840.1 GBGNSS/LiDAR/Camera/IMU/SPAN-CPTGoogleDriveDynamic Objects, high-speed dataset
<p align="center"> <img width="420pix" src="img/CAGoldenBridge20190828191451.png"> </p>

2. Hong Kong Dataset

2.1 Sensor Setups

The platform for data collection in Hong Kong is a Honda Fit. The platform is equipped with the following sensors:

<p align="center"> <img width="712pix" src="img/hongkong_sensor.png"> </p>

Intrinsic and Extrinsic Parameters, ROSBAG Information

The coordinates transformation between multiple sensors, and intrinsic measurements of camera can be found via Extrinsic Parameters and Intrinsic Parameters of Camera. The fish-eye camera intrinsic parameters can be found through here.

2.2. Dataset 1: HK-Data20190426-2

Brief: Dataset HK-Data20190426-2 is collected near Whampooa of Hong Kong.

Some key features are as follows:

Date of CollectionTotal SizeSensorsDownloadFeatures
2019/04/2641.6 GBGNSS/LiDAR/Fish-eye Camera/IMU/SPAN-CPTGoogleDriveDynamic Objects, Tall buildings
<p align="center"> <img width="420pix" src="img/HK-Data20190426-2.png"> </p>

2.3. Dataset 2: HK-Data20190426-1

Brief: Dataset HK-Data20190426-1 is collected near Ma Tau Kok of Hong Kong.

Some key features are as follows:

Date of CollectionTotal SizeSensorsDownloadFeatures
2019/04/2624.0 GBGNSS/LiDAR/Fish-eye Camera/IMU/SPAN-CPTGoogleDrivePoor GNSS visibilities, Very Tall buildings
<p align="center"> <img width="420pix" src="img/HK-Data20190426-1.png"> </p>

2.4. Dataset 3: HK-Data20190316-2

Brief: Dataset HK-Data20190316-2 is collected near Ma Tau Kok of Hong Kong.

Some key features are as follows:

Date of CollectionTotal SizeSensorsDownloadFeatures
2019/03/1662.3 GBGNSS/LiDAR/Fish-eye Camera/IMU/SPAN-CPTGoogleDrivePoor GNSS visibilities, Very Tall buildings
<p align="center"> <img width="420pix" src="img/HK-Data20190316-2.png"> </p>

2.5. Dataset 4: HK-Data20190316-1

Brief: Dataset HK-Data20190316-1 is collected near Ma Tau Kok of Hong Kong.

Some key features are as follows:

Date of CollectionTotal SizeSensorsDownloadFeatures
2019/03/1627.9 GBGNSS/LiDAR/Fish-eye Camera/IMU/SPAN-CPTGoogleDrivePoor GNSS visibilities, Very Tall buildings
<p align="center"> <img width="420pix" src="img/HK-Data20190316-1.png"> </p>

2.6. Dataset 5: HK-Data20190117

Brief: Dataset HK-Data20190117 is collected near Ma Tau Kok of Hong Kong.

Some key features are as follows:

Date of CollectionTotal SizeSensorsDownloadFeatures
2019/03/166.11 GBGNSS/LiDAR/Camera/IMU/SPAN-CPTGoogleDrivedecent GNSS visibilities, sub-urban
<p align="center"> <img width="420pix" src="img/HK-Data20190117.png"> </p>

3. Easy Use Scripts

3.1 Extract ground truth and u-blox solution to .kml file

cd ~/catkin_ws/src
git clone https://github.com/weisongwen/UrbanLoco
cd ../
catkin_make
source ~/catkin_ws/devel/setup.bash

3.2 Extract the raw GNSS measurements from /ublox_node/... to RINEX file

Some researchers may want to apply the RTKLIB to process the GNSS data using the RTKLIB which is mainly used in the GNSS field, we recommend to use one piece of code from ublox2rinex and issue.

3. Acknowledgements

The authors from Berkeley hereby thank the generous support of Robosense, whose donation of a Robosense R32 LIDAR is a critical step in our data acquisition process. We also thank Di Wang for his contributions on vehicle instrumentation at UC Berkeley.

4. License

This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License and is provided for non-commercial but academic use. If you are interested in using this dataset for commercial purposes, please contact us.