Using Neural Network to Control the Robot Arm (FK, IK, Trajectory Planning)

🎁 Get FREE Robotics & AI Resources (Guide, Textbooks, Courses, Resume Template, Code & Discounts) – Sign up via the pop-up at https://kevinwoodrobotics.com/
📝 Real-World Robotics Project Course Overview and Waitlist: https://forms.gle/xu4rfk7GWHCaZGBEA
🦾 Learn Robotics and AI: https://kevinwoodrobotics.com/product/all-courses-bundle-deal/
🤝 Consulting and Mentorship: https://kevinwoodrobotics.com/product-category/consulting/
👁️ Learn OpenCV: https://kevinwoodrobotics.com/product/opencv-python-tutorials-full-playlist/
🤖 Learn ROS: https://kevinwoodrobotics.com/product/ros2-tutorials-full-playlist/
🧠 Learn Computer Vision using AI: https://kevinwoodrobotics.com/product/computer-vision-using-ai/
🖥️ Learn AI and Machine Learning: https://kevinwoodrobotics.com/product/ai-and-machine-learning-bundle/
🛒 Amazon Store (My Top Picks for Robot Projects and More!): https://www.amazon.com/shop/kevinwoodroboticsai?ref_=cm_sw_r_cp_ud_aipsfshop_Q3AMRNBSVVB00Y1XR9TN_3

In this video, we're going to see if we can control a robot arm using a neural network! Is the robot going to have a mind of its own or will it follow the trajectory that we want? Let's find out!

I will be using a simple 2 DOF (degree of freedom) planar robot for this example, but the concepts can be expanded to n DOFs. I will start off by going over what forward kinematics is and how to derive the forward kinematics equations for a 2 DOF robot arm. Everything is verified visually using Python's matplotlib in VS Code, so I will cover how I am doing the robot animation as well. Then I will explain what is inverse kinematics and how to derive the inverse kinematics equations for a 2 DOF robot arm. Once we have that, we can do trajectory planning to control the robot to follow a desired path.

Now that we have covered the basics, we can try to do the same thing using a neural network. I will explain what a neural network is and how we can set up the problem. Using tensorflow, I will show you how you can construct the neural network model. Then I'll go over how I generated the training data for the inverse kinematics problem. Now, for the fun part, I will use the trained model to control the robot arm and compare it with different amount of training data. Lastly, I will discuss some things you can try.

Learn AI: https://kevinwoodrobotics.com/product-category/ai-tutorials/

0:00 Introduction
1:33 Virtual environment setup
2:01 What is forward kinematics?
3:51 Forward kinematics 2dof robot arm derivation
5:28 Forward kinematics python code
6:42 Robot animation python code
9:33 What is inverse kinematics?
10:17 Inverse kinematics 2dof robot arm derivation
14:26 Inverse kinematics python code
15:24 What is trajectory planning?
16:49 Trajectory planning python code
18:13 What is a neural network?
20:27 Neural network using tensorflow python code
21:08 Training data python code
22:55 AI robot control python code
25:07 Things you can try
Thanks for watching! If you found this video helpful, please like, subscribe and share:
https://www.youtube.com/@kevinwoodrobot

Social:
LinkedIn: https://www.linkedin.com/in/kevinwoodrobot
Github: https://github.com/kevinwoodrobot
Instagram: https://www.instagram.com/kevinwoodrobot/
Twitter: https://twitter.com/kevinwoodrobot