UrbanNav-SLAM

Simultaneous Localization and Mapping (SLAM) in crowded urban environments by fusing Lidar, IMU, GPS, Odometry, and SPAN-CPT data.

Kalman Filter details:

• Number of states tracked: 13
• (3 positions, 3 velocities, 3 accelerations and 4 quaternion orientations)
• State Vector: [x, y, z, vx, vy, vz, ax, ay, az, qx, qy, qz, qw]
  • Position states : (x, y, z)
  • Velocity States : (vx, vy, vz)
  • Acceleration States : (ax, ay, az)
  • Quaternion States: (qx, qy, qz, qw)

Data used:

UrbanNav Dataset:

UrbanNav is an Open-Sourcing Localization Dataset collected in Asian Urban Canyons, including Tokyo and Hong Kong. The dataset was developed to aid in solving the challenging problem of positioning and localization in deep urban canyons using low-cost sensors. The accuracy of GNSS is severely challenged in urban canyons due to the high-rising buildings, leading to numerous Non-line-of-sight (NLOS) receptions and multipath effects. Moreover, the excessive dynamic objects can also distort the performance of LiDAR and the camera. The dataset includes sensor measurements from GNSS receiver, LiDAR, camera, and IMU, together with accurate ground truth from SPAN-CPT system.

Hong Kong Dataset

1. Sensor Setups The platform for data collection was a Honda Fit. The platform was equipped with the following sensors:

   • 3D LiDAR sensor (HDL 32E Velodyne): (360 HFOV, +10~-30 VFOV, 80m range, 10Hz)
   • IMU (Xsens Mti 10, 100 Hz, AHRS)
   • GNSS receiver: u-blox M8T or u-blox F9P (to be updated)
   • camera:(1920X1200,79.4°X96.8°, 10Hz)
   • SPAN-CPT:(RTK GNSS/INS, RMSE: 5cm, 1Hz)

2. Dataset : UrbanNav-HK-Data20200314 Brief: Dataset UrbanNav-HK-Data2020314 is collected in a low-urbanization area in Kowloon which suitable for algorithmic verification and comparison. The coordinates transformation between multiple sensors, and intrinsic measurements of camera can be found via Extrinsic Parameters, IMU noise and Intrinsic Parameters of Camera.

Date of Collection	Total Size	Path length	Sensors
2020/03/14	27.0 GB	1.21 Km	LiDAR/Camera/IMU/SPAN-CPT

• Download by Dropbox Link:
  • UrbanNav-HK-Data20200314 (ROS)
    • ROSBAG file which includes:
      • 3D LiDAR point clouds: /velodyne_points
      • Camera: /camera/image_color
      • IMU: /imu/data
      • SPAN-CPT: /novatel_data/inspvax

Requirements:

• ROS Melodic
• A-LOAM - To process lidar points and output lidar mapping and odometry data
• Novatel Span Driver
• utm

Sensor topics used:

• /imu/data
• /aft_mapped_to_init
• /navsat/fix
• /novatel_data/inspvax

Steps:

1. Install ROS melodic from here, and utm.
2. Setup a ROS workspace with A-LOAM and Novatel Span Driver packages.
3. Download rosbag file for the UrbanNav dataset.
4. Run the following commands via terminal:

roslaunch aloam_velodyne aloam_velodyne_VLP_16.launch

rosbag play 2020-03-14-16-45-35.bag

python sensor_fusion_urbannav.py

5. The filtered odometry information of the vehicle can be accessed using

rostopic echo /ekf_odometry

Results:

Fused localization using LiDAR, GPS, SPAN data

X and Y position tracking using LiDAR, GPS, SPAN data

Fused localization using GPS and SPAN data only

X and Y position tracking using GPS and SPAN data only

Orientation tracking using LiDAR and IMU data

Limitations:

The filter was not tuned to predict position, velocity, and acceleration along the z-axis (altitude).