Week 4 (3rd Period)

Week 4 (3rd Period)

  • António Morais
  • 3rd Period - Weekly Progress
  • March 10, 2024

Website development

During the first week since the start of the 2nd semester and along with it, the ElectroCap Challenge, the team focused on launching the current website this is written on. As of now, it is still a very early version of the website and we already have feedback to start changing how it looks and it’s contents.

3D Mapping

→ A-LOAM

Thinking about Week 3 (3rd Period) problems encountered the team studied more in depth the algorithm and found that the anomaly stemmed from discrepancies in the algorithm’s code, particularly in the handling of .frame_id parameters, necessitating adjustments to accommodate variations across different versions of ROS.

// As an example
laserOdometry.header.frame_id = "/camera_init"; // ROS Melodic or Kinetic
laserOdometry.header.frame_id = "camera_init"; // ROS Noetic

With these modifications implemented throughout the codebase, the algorithm successfully processed the dataset (nsh_indoor_outdoor.bag), culminating in the generation of the subsequent 3D map:

alter-text
Figure 1 - Map obtained utilizing the A-LOAM algorithm alongside one of the datasets provided by the authors.

Subsequently, the team directed its efforts towards executing the algorithm within the TIAGo’s simulation environment developed by SocRob@Home. Easing this transition merely required supplying the algorithm with the data acquired by the Ouster LiDAR in simulation. To accomplish this, we identified the topic where the Ouster data is published:

// File path: A-LOAM/src/scanRegistration.cpp
ros::Subscriber subLaserCloud = nh.subscribe<sensor_msgs::PointCloud2>("/velodyne_points", 100, laserCloudHandler); // No changes
ros::Subscriber subLaserCloud = nh.subscribe<sensor_msgs::PointCloud2>("/ouster/points", 100, laserCloudHandler); // With changes

Following these adjustments, we successfully generated a map of the simulation environment:

alter-text
Figure 2 - Simulation environment (left); Map of the simulation environment obtained with A-LOAM (right).

Related Posts

Week 6 (3rd Period)

Week 6 (3rd Period)

  • Catarina Caramalho and Tiago Teixeira
  • 3rd Period - Weekly Progress

3D Localization When trying to simulate both algorithms, the students faced two problems.

Read More
Week 2 (3rd Period)

Week 2 (3rd Period)

  • António Morais
  • 3rd Period - Weekly Progress

3D Mapping → A-LOAM In the context of the Mapping module, our team’s objective was to deploy the A-LOAM algorithm, and subject it to comprehensive testing in a simulated environment.

Read More
Week 5 (3rd Period)

Week 5 (3rd Period)

  • António Morais, Catarina Caramalho and Tiago Teixeira
  • 3rd Period - Weekly Progress

3D Mapping → A-LOAM This week, the team collected real data from the robot’s Ouster LiDAR to conduct testing with A-LOAM .

Read More