Team 2152 -- Lidar/Jetson

2152 was mentioned in another thread about using Lidar and Jetson in auto this year. Does anybody here know how they implemented these two items together? We are interested in trying to using Lidar for field localization and we are looking for more information.

2152 uses Robot Operating System (ROS) for all of their non-roborio processing which consists of 3 Odroid X4U’s and one Jetson. The Jetson is running a neural network to detect cubes by object recognition, the Jetson sends an approximate angle and distance to Odroid12. Odroid12 publishes the approximate angle and distance and uses the LiDAR to find the angle, distance (centroid), and the closest point of cube and sends that data to the roboRio over UDP. The rio then uses 2 PID loops, one for angle and one for distance to acquire the cube. This is used in Teleop to assist the Drive Team in acquiring cubes. We got SLAM working but we never had enough time to use it in auto.

Command and Control handout :

Video of Robot autonomously picking up cubes:

Is there any chance your team will publish a whitepaper with the results you got running this system (accuracy, average fps, etc) ?

Thank you for posting the setup for running LIDAR with Jetson. We will delve more deeply into using the ROS system along with the LIDAR as we have seen same used in several non-FRC applications.

Hope to see a whitepaper detailing your system. The video and description have helped inspire our students to continue with their off-season plans.

Pardon my ignorance on this subject, but isn’t running 4 co-processors on a single robot a bit excessive? Surely this could have been consolidated somehow. The weight penalty alone for all that hardware would be an awfully hard sell, for me anyways. :ahh:

It depends on what each processor is doing. Being able to have a dedicated processor for a task does make development easier as well.

Plus, unless you were trying to double ramp this year making weight wasn’t too terrible and honestly, that seems to be a consistent theme lately.