I am doing a project for a machine learning course where I perform GNC with a turtlebot3, but utilize ML algorithms instead of the classic non-ML algorithms.

The objective is to reach an orange cone through an obstacle environment. The ML algorithms are:

Camera - objective detection - detect the cone's location - supervised learning - YOLO algorithm

LiDAR and Encoders - SLAM - detect obstacles and localize - unsupervised learning - unsure of ML algorithm

Encoders - Control - determine velocity command to travel along path - Reinforcement learning - PPO

So the gist of the plan is that using ML algorithms, the camera will detect the objective, the LiDAR and encoders will perform SLAM and the encoders and an A* path will determine the control commands.

I've got the turtlebot working so I next need to understand the ROS layout. I'm thinking my ROS graph will look something like the following. Does this look right? I am new-ish to ROS and so am trying to understand how this all works.

enter image description here



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.