top of page
James O. Vickery
I have a Bachelor's in Physics and Math and a Master's in Mechanical Engineering concentrating in Robotics. I made this website to showcase my various projects, particularly self-motivated passion projects. Explore some of my work below.
Quadruped Walking Robot
Project for Robotics Studio at Columbia University
"Bobo the Robo"
This project was for my Robotics Studio class. The goal of the class was to build a walking robot from scratch.
For the class, I was supplied with 8 motors, a serial to PWM controller, a battery, a voltage regulator, a Raspberry Pi, and a $100 budget for additional parts. Everything else is designed and built by me.
This project is a combination of mechanical design with electrical/motor integration, programming, simulation, and much more.
For this project I used both 3D printing and laser cutting.
For the class, I was supplied with 8 motors, a serial to PWM controller, a battery, a voltage regulator, a Raspberry Pi, and a $100 budget for additional parts. Everything else is designed and built by me.
This project is a combination of mechanical design with electrical/motor integration, programming, simulation, and much more.
For this project I used both 3D printing and laser cutting.
Current State of Walking
Just using motion primitives (simple sin and cos functions) I am able to achieve speeds of 19 cm/s while walking.
Unfortunately due to campus shutdowns in the spring of 2024 I was not able to implement some of the more advanced motion planning I discuss below.
As you can see in later slides I plan to improve upon this significantly by using simulation and reinforcement learning to optimize the walking gait for speed and stability.
Unfortunately due to campus shutdowns in the spring of 2024 I was not able to implement some of the more advanced motion planning I discuss below.
As you can see in later slides I plan to improve upon this significantly by using simulation and reinforcement learning to optimize the walking gait for speed and stability.
Initial CAD Design
This was my initial CAD design. Once I had this initial design, I analyzed it for potential problems. I successfully discovered a few potential problems and implemented them in the follow-on design.
Exploded View
Exploded view of the initial CAD Design
Final pre-build CAD
I modified the 3D printed lower legs due to potential issues discovered with initial CAD. This new bent design allows the robot to stand with much less torque being applied to the lower motors.
Bill of Materials
Automatically generated a bill of materials to determine the quantity of parts needed to be ordered and manufactured.
Initial Leg Testing
Testing the leg mechanism and motors. I did a preliminary test of the legs designs and got familiar with the python control library for these specific motors.
I used various table top testing instruments including Oscilloscopes, DMMs, and table top power supplies.
I used various table top testing instruments including Oscilloscopes, DMMs, and table top power supplies.
Initial Test Fit Mounting Legs to Frame
Assembly began with test fitting the legs to the frame.
Wires routed through holes in the leg to prevent being pulled loose and for better aesthetics.
Wires routed through holes in the leg to prevent being pulled loose and for better aesthetics.
Continuing Assembly
Connected all leg mounts and added in the laser cut mounting plate for the electronics.
Further Assembly
Almost all legs components mounted and properly connected.
Unpowered motors supporting weight
Demonstrating how the internal frictions of the motors prevent the robot from collapsing under its own weight, even when the motors are unpowered.
First Coordinated Movement of the Robot
First time testing the robot out a coordinated movement of the robot. This movement is to simply stand up from sitting.
Full Assembled Robot
All wires and cables connected and properly secured. Notice that the cables coming off the IMU are wrapped together to negate ground loop interference.
PyBullet Simulation
Successful simulation of the Robot in PyBulllet. I plan to use simulation and reinforcement learning to train the robot to walk.
Simulation
Here is an example of having the robot walk manually within the simulation using the same motion primitives as I did in real life.
Click image to see and read more
3 Degree of Freedom Manipulator
A self-inititiated project to help myself and others better understand topics in kinematics.
Realistic Render
I found it difficult to imagine certain robotics concepts when they were explained in very abstract terms.
My hope with this robot is that a professor can concretely demonstrate: Relative coordinate frames, forward/inverse kinematics, singularities, motion planning and trajectories.
I plan to mount a set of XYZ axes to the base and to the end of the arm to demonstrate, among other principles, relative coordinate frames.
My hope with this robot is that a professor can concretely demonstrate: Relative coordinate frames, forward/inverse kinematics, singularities, motion planning and trajectories.
I plan to mount a set of XYZ axes to the base and to the end of the arm to demonstrate, among other principles, relative coordinate frames.
Simplified Overview
The design is a 3-link elbow manipulator.
It uses 3 NEMA-17 Stepper motors actuate the joints
and is controlled by an Arduino microcontroller and an 3 a4988 stepper motor controllers. I'm hoping to upgrade these to better options (such as the TMC2208/2209) after initial testing.
Project has been designed, but assembly is stalled due to the laser cutters being upgraded. Hoping to have an assembled prototype in the near future.
After testing I hope to redesign the linkages to be more visually appealing.
It uses 3 NEMA-17 Stepper motors actuate the joints
and is controlled by an Arduino microcontroller and an 3 a4988 stepper motor controllers. I'm hoping to upgrade these to better options (such as the TMC2208/2209) after initial testing.
Project has been designed, but assembly is stalled due to the laser cutters being upgraded. Hoping to have an assembled prototype in the near future.
After testing I hope to redesign the linkages to be more visually appealing.
Exploded View
I used CAD to not only model my design, but also to determine interference issues and predict assembly issues. For example, I confirmed that it will be possible to attach the shaft collar to the shaft through the cutout on the base.
First Manufactured Part.
The 3D printed base with the base motor inside of it.
Simulating Stepper Motors
Using an Arduino simulator to develop the software for controlling the steppers. Current plan is to calculate the inverse kinematics in python and then use the serial bus to pass commands to the Arduino.
The code on the left successful takes in positional commands and moves the motors to the appropriate location.
The code on the left successful takes in positional commands and moves the motors to the appropriate location.
Click image to see and read more
Motion Planning for 2 DoF Planar Manipulator
Initially began as a homework assignment that I extended.
GUI Demonstrating Motion
This program dynamically simulates the movement of a 2-link planar manipulator.
This began as homework assignment: “Use MatLAB to simulate a straight-line trajectory between the points (1,0) and (0,2) with unit length links.” After submitting the assignment, I rewrote the program in python and modified it to include more capabilities.
The blue and orange lines are the two links and the green line is the path the path that the end effector follows. The Blue circle represents the workspace boundary.
This began as homework assignment: “Use MatLAB to simulate a straight-line trajectory between the points (1,0) and (0,2) with unit length links.” After submitting the assignment, I rewrote the program in python and modified it to include more capabilities.
The blue and orange lines are the two links and the green line is the path the path that the end effector follows. The Blue circle represents the workspace boundary.
Simulation
I extended the program so that it can now:
Simulate between any two points
Use any length of links
Use any number of interpolations
Show the workspace boundaries
Simulate between any two points
Use any length of links
Use any number of interpolations
Show the workspace boundaries
Click image to see more
Kelvin Probe Force Microscopy Modification Project
Haverford College Senior Thesis
Lab Overview
For my thesis I designed and assembled modifications to an Atomic Force Microscope (AFM) to perform multiple forms of Kelvin Probe Force Microscopy (KPFM).
I continued my predecessors work while introducing new elements and making design decisions.
I continued my predecessors work while introducing new elements and making design decisions.
Principles of AFM
Atomic Force Microscopy (AFM) works by using contact* forces and feedback to detect the topography of a sample. A microscopic tip essentially taps along the surface of a sample and maps out the surface by feel.
*In reality it is electrostatic forces but contact is a good enough analogy.
Image taken from: Reilly, Simon & Leach, Richard & Cuenat, Alexandre & Awan, Shakil & Lowe, Martin. (2006). Overview of MEMS sensors and the metrology requirements for their manufacture.
*In reality it is electrostatic forces but contact is a good enough analogy.
Image taken from: Reilly, Simon & Leach, Richard & Cuenat, Alexandre & Awan, Shakil & Lowe, Martin. (2006). Overview of MEMS sensors and the metrology requirements for their manufacture.
Principles of KPFM
Kelvin Probe Force Microscopy (KPFM) is a modification of a AFM that allows detection of electrical signals in addition to topography.
This is generally accomplished by making multiple passes over a sample. The first pass detects topography and then the second pass follows the same path but at a fixed height above the sample. If the correct voltage is applied to the tip it allows for measurement of the localized work function of the sample.
Image taken from: https://blog.brukerafmprobes.com/guide-to-spm-and-afm-modes/liftmode/
This is generally accomplished by making multiple passes over a sample. The first pass detects topography and then the second pass follows the same path but at a fixed height above the sample. If the correct voltage is applied to the tip it allows for measurement of the localized work function of the sample.
Image taken from: https://blog.brukerafmprobes.com/guide-to-spm-and-afm-modes/liftmode/
KPFM Results
Image scale: 5 microns x 5 microns.
Left image (topography): 120nm from highest points (gold) to lowest points (brown)
Right image (phase/voltage): 0V (gold) to 1.5V (brown).
The left image is an AFM image of the patterned substrate.
The golden colored vertical lines are conductive gold-titanium “fingers."
The right image shows the KPFM system correctly detecting the difference in electrical charge applied to alternating fingers.
The jagged randomly oriented streaks across the image are porphyrin nanowires.
Left image (topography): 120nm from highest points (gold) to lowest points (brown)
Right image (phase/voltage): 0V (gold) to 1.5V (brown).
The left image is an AFM image of the patterned substrate.
The golden colored vertical lines are conductive gold-titanium “fingers."
The right image shows the KPFM system correctly detecting the difference in electrical charge applied to alternating fingers.
The jagged randomly oriented streaks across the image are porphyrin nanowires.
Feedback Computer and Driving Circuit.
The computer and circuit shown drove the electrical feedback necessary to perform KPFM.
Soldered Circuit
I moved the original solderless breadboard based circuit to a more permanent soldered breadboard and made efficiency improvements that decreased complexity.
Atomic Force Microscope
This is the modified AFM for the projection.
I had to become intimately familiar with both proprietary software and working principles of operation.
I had to become intimately familiar with both proprietary software and working principles of operation.
Click image to see more
Timing Chain Replacement
(1996 Chevy S-10)
Hands-on experience
Timing Chain Replacement
I have a passion for cars and during my summer break one year I purchased a 1996 Chevy S-10 with a loose timing chain for $400. I then spent three days replacing the timing chain on the 2.2L 4-cylinder by myself. (The old timing chain is in my hand)
The '96 S-10 had the ability to replace the timing chain without removing the front end or taking the engine out of the car which is why I took the project on.
The '96 S-10 had the ability to replace the timing chain without removing the front end or taking the engine out of the car which is why I took the project on.
Reassembled engine bay after Testing.
After reassembling the engine, to my pure delight the truck started right up.
I had been nervous about the timing since the 2.2L L4 engine was an interference engine.
I had been nervous about the timing since the 2.2L L4 engine was an interference engine.
Truck after Test Driving
In hindsight I regret not taking more photos of the project, but we deal with the mistakes of our past selves.
A few weeks later, I replaced the brake calipers and pads, but unfortunately the ABS system was broken and it was impossible to get a replacement unit.
A few weeks later, I replaced the brake calipers and pads, but unfortunately the ABS system was broken and it was impossible to get a replacement unit.
Click image to see more
Python and Primes
A few fun scripts for factoring and generating primes/semi-primes
Programs with Primes
The idea came from reading about Shor’s algorithm/RSA encryption. I decided to practice python by making programs with various tasks
You can see the programs on the left generating primes and semi-primes while the programs on the right confirm the numbers on the left and demonstrate further capabilities.
Further info on GitHub:
https://github.com/JOVickery1/Primes
You can see the programs on the left generating primes and semi-primes while the programs on the right confirm the numbers on the left and demonstrate further capabilities.
Further info on GitHub:
https://github.com/JOVickery1/Primes
Click image to see more
Model of a Piston Engine
An educational tool for fellow students.
Simplified Working Model
I used this model to explain working principle of an internal combustion engine to fellow classmates during high school.
Functional 3D model was printed. Explanation of the 4-stroke engine cycle laser engraved into the wooden base.
Functional 3D model was printed. Explanation of the 4-stroke engine cycle laser engraved into the wooden base.
Original CAD
Original CAD model. Unfortunately the files were lost during a computer transfer.
Click image to see more
Connect
bottom of page