Train Deep-Learning-Based CHOMP Optimizer for Motion Planning Obstacle Setting

Dear MATLAB users.

Hello.

I am running Train Deep-Learning-Based CHOMP Optimizer for Motion Planning example

https://mathworks.com/help/robotics/ug/train-deep-learning-based-chomp-optimizer.html

In the example I want to change the positions of the obstacles, and try it again.

However, I cannot find the section of the code to modify for the purpose.

If you can find what and where I need to change, please let me know.

Thank you.

Answer 1

[Aravind]

Hi @Choonghan,

If you want to change the positions of obstacles and test the path planning of the trained “dlCHOMP” path planner in a different target environment, you can use the “generateSamples” function. You can find detailed steps in the “Infer Using Trained dlCHOMP Optimizer” section of the example here: https://www.mathworks.com/help/robotics/ug/train-deep-learning-based-chomp-optimizer.html#GettingStartedWithDLCHOMPExample-6.

By adjusting the “Seed” value in the data options, you can create various target environments. Each distinct “Seed” value will generate a different environment for your tests. For further details on the “generateSamples” function, you can check the documentation at: https://www.mathworks.com/help/robotics/ref/dlchomp.generatesamples.html.

I hope this helps address your question.