Making an 8DOF Quadruped
I tried my hand at making a Quadruped Robot; an 8-DOF one as opposed to the typical 12-DOF ones. It is still very much a WIP, but I just couldn’t wait to make a post about it.
In the top most photo you can see how it is taking serial inputs from a GUI application. The application itself was made using Processing. The GIF below shows how I send the servo values using the interactive knobs and pressing the respective servo buttons. For instance, HLF would mean Hip Left Front and KLB would be Knee Left Back. So if I set the HLF knob to be of a certain value, and then press the HLF button, it will send a command to move the front left servo connected to the hip, to the desired position. The GUI establishes Serial communication with the Arduino.
Spent so many hours 3D Printing the body. (The STL files were obtained from this Instructables page). To provide sufficient current to all my servos, I also bought a 5V 5A DC Switching Power supply. Which you can see in the GIF above, the silvery box with many wires coming out of it.
I had scaled down my project to 8DOF because I want to start all the inverse kinematic analyses on that first before moving on to a 12DOF Quadruped. (Which is why you can see the Hip servo rig isn’t placed here as opposed to the original model.)
For now, you can see me feeding the joint angles to the robot and observing the motion. Here, you can see it lie down fully, then attempt a push up. Well, its version of a pushup!
Please excuse the mess behind 🙇 |
Also, this may be slightly off topic but I find it funny how most robotics textbooks haven’t updated their curriculums since the 80s. We have more robotic configurations, like quadrupeds, hexapods, etc... It would have been really helpful to see examples of Quadruped’s kinematics and how to find its Jacobian derivatives in textbooks. For now, you have to search for research papers or hope to find someone’s publicly available PhD thesis to understand how it’s done.
So eventually figuring out the Inverse Kinematics and trajectory planning would be next steps for this project. Though it might take a while before I get started on that... I always like to joke that solving a robot’s kinematics (and that too its inverse kinematics) is just as baffling as attempting to bake a perfectly-shaped éclair. Of course one of them is harder... the eclairs, that is. 😝
Waving you goodbye! |