Picked from the software, topic, and level of “Class Reinforcement Learning Robot Manipulator #1: Forward and Inverse Kinematics of a 2-DOF Robotic Arm”.
At most one email a week. Unfollow anytime with one click.
Transcript
The full narration of this video. The current sentence is highlighted while the video plays; click any sentence to jump there.
Show transcript
Introduction
This is the final result, a two-joint robotic arm whose position we calculated using forward kinematics, and whose joint angles we calculated using inverse kinematics, complete with workspace and circular path that will be tracked throughout this series. Let's get started. Hello, welcome to the research tutorial series, adaptive control based on reinforcement learning for robotic manipulators. Industrial robotic arms must remain accurate even when carrying varying loads or being disturbed. In this 6-part series, we build everything from scratch, kinematics, dynamics, PID control and competitors, reinforcement learning environment, soft actor-critic agent training, up to adaptive testing against load and disturbances. In the first episode, we start with kinematics.
Step 1: Installing the library
We open the terminal through the terminal menu, new terminal, then install numpy, matplotlib, tork, gym for the reinforcement learning environment, and stable baseline 3 which contains the SAC algorithm.
Step 1: library
All our robot models are written from scratch without a dedicated simulator.
Create a file named robot.py. The arm has two links, each half a meter and two kilograms. Forward kinematics calculates the elbow and end-effector position from the two joint angles using cosine and sine. Inverse kinematics does the opposite. From the target point, the second joint angle is obtained using the cosine law, then the first joint angle. There are two solutions, upper elbow and lower elbow, and if the point is too far, the function reports the point as out of reach. The path function creates a circular path with a radius of 20 cm and a period of 4 seconds, then converts it into joint angles, velocities, and accelerations. Create a file named test_underscore_kinematics.py and run it.
Step 3: Two inverse kinematics solutions
Angle 00 produces the end-effector at 1 meter, arm straight horizontally. Angle 90 0 produces 0, 1, arm vertical. For point 0,6, 0,3, there are two solutions. Elbow upper around 74 and minus 96 degrees, and elbow lower around minus 21 and 96 degrees. Both of them we check back with forward kinematics and exactly return to the initial point.
Step 4: Workspace & arm pose
Let's see the graph. Thousands of blue points are the workspace, all points that can be reached by the end-effector, forming a circle with a radius of 1 meter. Green and orange lines are two configurations for the same point. The right graph shows the arm pose along the red circular path, the path that will become the task for all controllers in this series. Done.
Done!
We have already written forward and inverse kinematics of a two-joint robotic arm, understood the upper and lower elbow solutions, viewed the workspace, and created a circular path. In the next episode we will model its dynamics, forces and torques that drive the arm. Currently working on a thesis, dissertation, or research in the field of robotics and reinforcement learning? Cerdas Inovasi Digital helps implement robotic control, from manipulator kinematics and dynamics, PID and competitors, gym environment, up to Soft Actor Critic agent training, as well as opening Python programming courses for electrical and mechatronics engineering. Want to master the entire series, from robotic arm kinematics to adaptive reinforcement learning agent testing? All six episodes, complete with step-by-step PDF guides and source code, are available to premium tutorial subscribers. Subscribe starting at Rp69.000 per month, save 28% with the 12-month package, or one-time payment of Rp999.000 for lifetime access. Want just one class? Buy single class with lifetime access. Subscribe now on the premium page at cerdasinovasidigital.com. Contact our WhatsApp, or visit cerdasinovasidigital.com. Thank you for watching, see you in the next episode.
Ask the class material
Type a question or keyword; we find the part of this episode that explains it and jump the video there.
My notes
Notes are saved with the current video time and are only visible to you.
Questions & answers
Stuck on a step? Ask here; the instructor and other learners can answer.
No questions yet.
Your instructor
Cerdas Inovasi Digital Team
Software developers, BIM & 3D engineers, and AI practitioners
53 video classes1,290 episodes
We build software, BIM models, games and AI automations for clients every day, and turn that hands-on work into step-by-step classes: every step is shown on screen and comes with a PDF guide you can keep.
