Deep RL Arm Manipulation
Basic Requirements
Criteria | Meet Specification |
---|---|
Include in your project submission your write-up (PDF format), supporting images or videos, and the entire project folder minus the build folder. |
The student submitted all required files specified in the criteria. |
Objectives
Criteria | Meet Specification |
---|---|
Have any part of the robot arm touch the object of interest, with at least a 90% accuracy for a minimum of 100 runs. |
The student should complete all tasks specified in the Classroom, with the end objective of the robot arm touching the object with at least a 90% accuracy for a minimum of 100 runs. |
Have only the gripper base of the robot arm touch the object, with at least a 80% accuracy for a minimum of 100 runs. |
The student should complete all tasks specified in the Classroom, with the end objective of the arm's gripper base touching the object with at least a 80% accuracy for a minimum of 100 runs. |
Writeup Requirements
Criteria | Meet Specification |
---|---|
Provide a Writeup / README that includes all the rubric points and how you addressed each one. You should submit your writeup as pdf. |
Student includes a full write-up covering the required sections with supporting images where appropriate. The write-up must be submitted in PDF format. |
Reward Functions: Explain the reward functions that you created. |
Brief explanation of each reward function and associated reward values. The writeup should also include what type of joint control was implemented. |
Hyperparameters: Specify the hyperparameters that you selected for each objective, and explain the reasoning behind the selection. |
Student should explain the choice of hyperparameters for both objectives. |
Results: Explain the results obtained for both objectives. Include discussion on the DQN agent's performance for both objectives. Include watermarked images, or videos of your results. |
Student should describe and briefly explain the results they achieved for both objectives. The discussion should also include their comments on the DQN agent's performance and if there were any shortcomings. Student should include either watermarked images of their results, or attach a video that displays the results and the arm in action. |
Future Work: Briefly discuss how you can improve your current results. |
Student should discuss on what approaches they could take to improve their results. |
Tips to make your project standout:
Student attempted the additional project challenges and got good results for each of those, or student made significant changes to the project that indicates a good deal of well thought out experimentation in order to improve accuracy - for example, defining complex reward functions based off of research papers etc.