Sensor Suite ============ Interchagable Payload Unit -------------------------- Sensing suite is desinged as an interchagable playload unit between **Bell 412 Advanced Research Systems Aircraft (ASRA)** and **DJI M600 hexacopter drone** platforms. The payload unit is capable of running visual-inertial-LiDAR odometery and mapping algorithms real-time while collecting data. .. image:: /images/payload.jpg :width: 50% .. image:: /images/platforms.jpg :width: 49% Sensors ------- .. .. figure:: /images/platforms.jpg .. :width: 99% .. :align: center .. :alt: Caption for first image .. The payload is interchagably mounted between **Bell 412 Advanced Research Systems Aircraft (ASRA)** .. and a **DJI M600 hexacopter drone**. IMU +++++++++++++++++++++++++++++++ The IMU publish the following ROS topics. .. csv-table:: :header: Message Type, Topic Name, Rate [Hz], Description :widths: 20,20,20,20 sensor msgs/Imu, /imu/data, 400, raw data sensor msgs/Imu, /imu/data stamped, 400, raw data with trigger hardware timestamp geometry msgs/Vector3Stamped, /imu/mag, 100, raw magnetometer data sensor msgs/TimeReference, /imu/time_ref, 400, hardware timestamp at IMU trigger sensor msgs/TimeReference, /imu/time_ref_cam, 20, hardware timestamp at camera trigger sensor msgs/TimeReference, /imu/time_ref_pps, 1, hardware timestamp at PPS receive Cameras ++++++++ The two cameras are publising the following ROS topics. .. csv-table:: :header: Message Type, Topic Name, Rate [Hz], Description :widths: 20,20,20,20 sensor msgs/Image, /camera/image_mono, 20, Mono8 format down camera image sensor msgs/Image, /camera/image_color, 20 , RGB8 format down camera image sensor msgs/Image, /front_camera/image_mono, 20, Mono8 format front camera image sensor msgs/Image, /front_camera/image_color, 20, RGB8 format front camera image LiDAR ++++++++ LiDAR publish the following ROS topics. .. csv-table:: :header: Message Type, Topic Name, Rate [Hz], Description :widths: 20,20,20,20 sensor msgs/PointCloud2, /velodyne_points, 10Hz, raw velodyne pointcloud sensor msgs/TimeReference, /time_ref_scan, 10Hz, hardware timestamp at pcd trigger GNSS Receriver ++++++++++++++ GNSS module publishes the following ROS topics. .. csv-table:: :header: Message Type, Topic Name, Rate [Hz], Description :widths: 20,20,20,20 sensor msgs/NavSatFix, /fix, 5, raw RTK-GNSS sensor msgs/NavSatFix, /fix_ppk, 5 , PPK GNSS nmea msgs/Sentence, /nmea_sentence, 5, nmea formatted GNSS nav msgs/Odometry, /fix_frl, 5, 6DOF pose ground-truth GNSS Base Station +++++++++++++++++ Reach RS2 Based station is used during data capture. The real-time NMEA stream for the simpleRTK2B was provided by Reach RS2 base station. RS2 communicates with the ZED-F9P using a ultra high frequncy (UHF) LoRa radio. The Reach RS2 also provides a post processed kinematics (PPK) solution upto 100km by logging data at 10Hz. Jetson Xavier AGX GPU --------------------- The Xavier is running on Ubuntu 18.04 and NVIDIA L4T 32.3.1 with Cuda 10.1 support, and ROS Melodic is installed on it. Additionally, the Xavier board is equipped with a 1GB SSD to store the ROS bags and a Wi-Fi module for data visualization purposes. Hardware Time-Synchronization Scheme ------------------------------------ The IMU publishes a :math:`/imu/time\_ref` topic, including the hardware timestamp of the IMU's internal clock for the published data. The point cloud acquisition hardware time and the camera trigger pulse hardware timestamp are pub- lished on :math:`/time\_ref\_scan`` and :math:`/imu/time\_ref\_cam`` topics, respectively. The timestamp of the received PPS message from the GNSS receiver is published on the :math:`/imu/time\_ref\_pps`` topic. These hardware timestamps can be used to accurately time-sync sensor messages. The block diagram of the hardware synchronization is shown below. .. image:: /images/sync.svg :width: 50%