Some image

Autonomous Target

Striking Boat

2.017 - Design of Electromechanical Robotic Systems

Spring 2019

Designed, built, and tested an autonomous model-sized boat on a team of 3. The boat was designed to accomplish a set of progressing goals. First, a propulsion and steering system that could switch between computer and manual control was constructed using individually actuated double propellers, a paired double rudder, an Arduino Mega for computer control, and an RC system for manual control. Next, to develop a more robust autonomous system, fault protection for flooding, overheating, and collisions was implemented using relevant sensors and the Arduino for control. Last, a striking mechanism was implemented using a Raspberry Pi 3 for computation, a webcam for vision, a servo based pan-tilt mechanism for aiming, and an automatic Nerf ™ gun for shooting. The vehicle successfully demonstrated fault protection functionality, was able to semi-autonomously locate and shoot marked targets, and was able to autonomously traverse a pre-decided path in a controlled test environment.

I was the electronics lead for our team and additionally worked on our target tracking/striking abilities. As a team, we also wrote a project proposal and a project report, and presented our project. Below is a photo of the boat in a tank as we tested it, featuring a lot of duct tape.

boat photo
Some image

Modelling the Elastic

Stiffness of a Discretely

Assembled Tower

2.671 - Measurement and Instrumentation

Spring 2019

2.671 centers around individual student projects where we design and conduct our own experiments. I looked at the elastic stiffness of the compenent parts and of an assembly of a tower I designed. My research paper can be found here, and my poster can be found here.

If you don't want to read the above, my project abstract follows:

The production of complex objects with specific electro-mechanical properties can be made easier by using discretely assembled structures, i.e. structures made out of repeating, individual components. Current research on discretely assembly structures is typically limited to highly technical applications, though discrete assemblies can be especially useful in less specialized applications because of their simplicity to produce. To simplify the design process of these structures, a simple spring model was used to predict the elastic stiffness of a discretely assembled tower. The same geometries were tested under compression for four materials: stainless steel, medium density fiberboard, cast acrylic, and corrugated fiberboard. The actual stiffness of the tower assembly was then compared to the modelled stiffness of the tower. The modelled stiffness overlaps with the measured stiffness for MDF and acrylic, indicating potential use for materials in this stiffness range, while it is unable to predict the stiffness of steel and cardboard.

tower photo tower photo
Some image

Smart Facades to

Reduce Residential

Building Energy


Undergraduate Research for Prof. Larry Sass's Design Fabrication Group

Summer 2018

For the first half of the semester, I explored using the hygroscopic properties of wood to create self-ventilating facades. I tried a variety of wood types and thicknesses in a bilayer configuration to see how effectively a curling behavior could be programmed into the wood.

For the latter half of the summer, I worked on developing a piezoelectric roof tile that could convert the kinetic energy of rain/snow/hail into electrical energy. The concept was to use a flexible PVDF strip covered in a weather-proofed fabric over an inclined rigid tile. The visual appearance of the tile would be similar to traditional shingling, and the geometry of the underlying rigid tile would help maximize the distortion experience by the piezolectric element.

some image

photo of a hygroscopic curling test

some image

a full size tile backing prototype

some image

close-up of a works like 1/8" scale model of the tiling system

Some image

Electron and Ion

Energy Analyzer

Lab Assistant at UCLA Plasma Physics Lab

Summer 2017

The Lab wanted an electron and ion energy analyzer for a cold plasma, that could be used for their undergraduate research class. The energy analyzer needed to be cheap and reliable to use. I designed and prototyped a new electron and ion energy analyzer based on existing literature. The analyzer was designed to be usable for either electron or ion detection, in an ambient plasma.

The analyzer consisted of an outer aluminum casing that I milled, with a gold plate with a 50 micron entrance hole. Three grids, separated w/ mica spacers, behind that were used to filter for the desired particle/energy range, and a ions or electrons would deposit on a final copper plate. The filter grids and copper plate went to a very, very simple PCB that I printed, and wires

Some image

The complete analyzer, unnattached to a probe.

Some image

The analyzer with its cap removed, exposing the first grid.

Some image

The analyzer attached to a probe

Some image

The analyzer being tested in a plasma.

Some image

Comparative B-Dot

Probe Study

Lab Assistant at UCLA Plasma Physics Lab

Spring 2017

B-dot probes are a ubiquitous diagnostic tool, though their design is frequently greatly varied without much evidence for why the design changes are beneficial. To that end, I worked on doing a comparative characterization of most common b-dot probe designs. Below is a draft abstract for the project:

Utility of several magnetic pickup loops used in plasma measurements are assessed from the point of view of the ratio of magnetic signal to electrostatic noise. The electrostatic pickup and the magnetic field frequency response are measured for a variety of probe geometries described in the literature. The electrostatic pickup is tested by placing the probe in an approximately uniform electric field generated by a parallel plate capacitor, a Faraday cup, and a split-Faraday cup. The magnetic field frequency response is determined by creating known magnetic fields using both a small single turn exciter and a Helmholtz coil. The electrostatic pickup and magnetic frequency response of the probes are tested from 10KHz to 25MHz using an Agilent E5100 Network Analyzer to measure the response, and plots are generated of the ratio of the measured signals from the electrostatic or B-field excitation test sources.

Some image

3-axis coaxial b-dot

Some image

another 3-axis coaxial b-dot

Some image

an in-progress 3-axis coaxial b-dot

Some image

a 3-axis twisted pair B-dot

Some image

an admittedly poorly lit photo of most of the b-dots tested for this project

Some image

Class Portfolio for

How To Make

(Almost) Anything

MAS.863 - How to Make (Almost) Anything

Fall 2018

How to make (almost) anything (MAS.863/4.140/6.943) is a great class taught by Prof. Neil Gershenfeld. The class is structured so that each week, you complete a different project which teaches a fabrication technique, and at the end of the semester, you complete a final project incorporating most of those skills.

For the class, we maintained a website that documents our weekly projects. You can find my website here.

Here are some photos of various projects I completed or collaborated on.

CNC mill made in 1 week as a group project

Some image


I make inflatables as a hobby and sometimes for class project(s). here are photos of them!

Some image

small(er) projects

Variety of projects

This is an autonomous ball collecting robot I made on a team of 5 for the 2018 MIT Mobile Autonomous Robot Lab competition. It's named humphrey. This is the CAD for a loft I made in my dorm room. The following is an in-construction photo of it. And completed, cropped to avoid clutter: This is a hot-spot healing sleave for dogs I worked on in 2.00 Introduction to Design with a team of 4. The image is a demo of the works-like model. The following are the looks-like prototypes.These are some semi-failed probes I made at the plasma lab.

Some image



Maybe everything else here is classifiable as either art/dumb/fun too, I don't know.