Software

RADAR

Detection

  • Begin detecting obstacles within regions of interest
  • Adjust speed constraints as obstacle approaches stopping boundary

Allocation

  • Software assigns points to sectors on each axis
  • Can find closest obstacles in each dimension without extra work

Avoidance

  • Based on the closest obstacle, we compute an angle tangent to it as a goal
  • We iteratively modify the users input to change the angle of approach

pip install pymmWave

  • Published an open-source Python package for wide-usage
  • Published optimized sensor interaction tools
  • Provided a set of complex custom RADAR algorithms for pose estimation, continuous state estimation, and more with a single RADAR sensor.

LiDAR

  • Two packages that we got from GitHub

  1. rplidar

  2. hector_slam

  • Changed a few parameters to perform scan matching and pose estimates as defined by Hector SLAM (Simultaneous Localization And Mapping)

  • Using ROS (Robot Operating System) we created a ROS node that subscribes to topics which is an outlet for the data collected by the LiDAR. So every time there is new data we process it by doing transformations using translation and rotation data along with point cloud data to determine the robot's position and proximity to its surroundings.

  • Used data in addition to RADAR data to see behind the vehicle

Software Flow