ROBOTICS CAPSTONE PROJECT
I am working with a team of 6 people to create two robots which are used in a robotics competition for a 20-week senior capstone project. The challenge involves designing two RC-controlled robotic cars capable of picking up 3D-printed pellets and ping pong balls and summiting a steel-coated pyramid with an incline angle of 37° to deposit pellets into a small hole under a tight budget. To make the competition more interesting, the robots are allowed to engage in combat with one another and have to be autonomous for the first 30 seconds of each round of the competition. We spent the first three weeks of the project designing the robot, determining the necessary electronics, and reviewing our designs in both a CDR and PDR with the capstone professor and TAs.
In the first 10 weeks, I worked with one other teammate on designing the electronics for the two robots. We selected the motors based on a torque and max speed which worked best for our game strategy. We used two 2×30 A RoboClaw motor controllers per robot and picked a switch and fuse to prevent damage to our RoboClaws. We determined the best performed modular testing of the RoboClaws and motors, RC controller, and Arduino to better understand how to sync the controller controls with driving, intaking, and depositing pellets. This required configuring the settings on the RC controller with the correct connections from the receiver to the Roboclaws and the eventual programming of the Arduino. By the end of the first half of the project, we were able to drive the robot, navigate it around cones, and drive up and down the pyramid.
In the second 10 weeks of the quarter, I worked on redesigning the robot in SolidWorks, specifically the drivetrain and upper assembly which houses our robots’ intakes, prototyping and machining the robot, and testing it. After our first demonstration at the end of the first 10 weeks, we iterated our design to feature longer, more protective walls to accommodate for the combative element of the competition. I took responsibility over managing our CAD and created separate assemblies for each of our robots and frequently 3D printed parts to prototype subassemblies and conduct tests of the intake mechanism. I worked with the drivetrain subteam to machine the drivetrain walls, cut churros and shafts, design an improved gathering and intake mechanism, and perform tests. Since the inclined surface of the pyramid is made from steel, we also designed magnet mounts with a low clearance to the ground that increases the robot’s effective weight only during climbing periods.
Most recently, we completed another demo to test our ability to collect pellets and ping pong balls, climb to the top of the pyramid, deposit pellets at the top of the pyramid, and descend down the steep brachistochrome.
AMATEUR LIQUID ROCKET
I led the engineering team to design, build, and test a bi-propellant liquid rocket. This was part of a group of amateur rocket-hobbyists called Project Caelus. Over the course of three years, we designed an entire engine, including the CAD of the injector, chamber, and nozzle using Fusion 360, the Piping & Instrumentation Diagram (P&ID), and the recovery system. We fabricated the injector and nozzle using a CNC machine, vertical drill press. I particularly took charge of the injector and nozzle design and testing, but also worked extensively on the plumbing including the design of the P&ID, assembly, and leak testing. We assembled a test stand to house the plumbing and ran multiple successful water cold flows of both propellant towers separately and together. We raised over $10,000 in funding and worked with NASA aerospace engineers and Aerojet Rocketdyne to complete PDR and CDRs of our rocket design.
Engine specifications:
– 2.5 kN thrust
– nitrous oxide & ethanol propellants
– self-impinging doublet injector
– deLaval nozzle
RATATOUILLE-INSPIRED ELECTRONIC HEADBAND
I worked with a two other Caltech students to create an electronically-controlled headband inspired by the movie Ratatouille. In the movie, a rat sits atop a chef’s head and pulls the chef’s hair to control his motions. I taught myself how to use Blender for this project and used it to make a model of the rat from scratch, which we 3D printed, painted, and secured to a headband. We programmed servo motors to move randomly using an Arduino Nano Every and attached them to the rat’s arms. The user’s hair can be threaded through the rat’s hands, which were designed to be closed loops to prevent hair from falling out. We also integrated two red LEDs for the eyes for humor. We documented this project through its entirety using Hackaday. Our project was selected to be presented at the LA Maker Faire + City of STEM event at USC, which included over 20,000 attendees.
GEARBOX + competition
In this project, I worked with a team of 4 others to design and machine a gearbox capable of turning a bike wheel. We were given a motor and test bed, as well as a tight budget to purchase acrylic, gears, and shafts. Under these constraints, we were challenged to create a gearbox which would increase the motor’s torque and speed as much as possible in a competition against other students taking the class. We opted for plastic gears due to their cheap budget and weight reduction, and calculated a gear ratio based on our desired output speed and torque. We created a CAD of the gearbox in SolidWorks and machined the gearbox using a manual mill, lathe, and various other tools in Caltech’s machine shop. At the end of the competition, we ranked second overall, with the highest maximum speed achieved out of all of the teams.
solar flare data visualization tool
I worked with a smaller version of Caltech’s Data to Discovery Program, where computer scientists and user design students are brought together to analyze a very large dataset that a research group at Caltech and/or JPL need assistance visualizing in order to find trends to inform research that would otherwise be impossible. I worked with Dr. Alan Labrador and Dr. Ashish Mahabal from the Cahill Center of Astrophysics at Caltech to design a web tool which visualizes solar flare data over the course of 7 years. The dataset included particle flux data over time steps for up to 15 elements across 8 energy bands and included data from both the STEREO and ACE solar probe missions. I reduced the dataset from a scale of ~1 TB to a few MB by creating summary JSON files which only extracted the parts of each solar flare event which were the most useful for the scientists’ research. Each solar flare event is represented by a circle with radiating rectangles/wedges which is intended to both artistically resemble the sun’s corona, where solar flares are visible, and create a visual pattern for the user to associate each element with. The length of each bar is scaled logarithmically to the particle flux intensity across a solar flare event peak, with each energy band stacked to still preserve the variation in particle flux intensity across energy bands. The tool allows users to easily toggle on and off elements to visualize, as well as sort by various features including the event date, event duration, and element ratios.
I created this tool using JavaScript, HTML, Python, and CSS. I conducted user interviews and workflow reviews with the researchers and used p5.js to prototype the design. Throughout the entire project, I worked with the researchers to ensure that the visualizations were accurately representing the data in the way they intended. At the end of this report, I wrote a research paper which won the Joel and Marcella Bonsall SURF Technical Writing Prize. This tool is planned to be implemented as a part of NASA’s official ACE and STEREO mission websites and is continued to be used by Caltech’s heliophysics research group.
galvo-based fiber switcher
I worked with the Hutzler Group at Caltech to create a single-mode fiber switcher for precision measurement physics applications. One of the design-aspects of this project featured designing a casing for a high-voltage power supply unit which I created in CAD using OnShape and 3D printed. I soldered and assembled the box to contain the electronics and later performed tests which validated the power supply unit’s function.