Donovan Crowley

×

Our team won 1st place in the 2025 Vertical Flight Society Design-Build-Vertical-Flight Competition

Background

This project involved building a drone to compete for the VFS Competition, focusing on agility, speed, payload capacity, precise payload dropping, and autonomous capabilities.

Process

First, I developed an algorithm using OpenCV's Hough Circle Transform to detect circular edges in the images from the ESP32 camera. My subteam tested with the microcontroller and used a servo to drop specific payloads on the desired targets.

Next, I worked with a Raspberry Pi Model B and configured it as a universal access point using Tailscale VPN and a Panda Wireless USB Adapter. This was a bit of a challenge and took quite some time to figure out how to do.

As this was my first time with the Raspberry Pi, I had to not only figure out how to set it up but also work around the school WiFi system. I would often have access timeouts using the Raspberry Pi operting system on the school WiFi, and I had to solve this issue through power cycles, flashing the SD card, or disconnecting and reconnecting the Ethernet cable before I finally configured the Pi on the USB. This hardship led me to learn about Wireless LAN priority, virtual environments, and operating systems.

I have since used the Raspberry Pi to control the electromagnet of the drone and detect distinct ArUco markers to operate its autonomous tasks. Currently, I'm developing the drone's LiDAR system through the I2C protocol to prepare for our upcoming competition this Spring 2026.

I also soldered the components shown in the schematic below to assemble six PCBs, connectioning the Pixhawk with each of the six motors on the drone.

×

Our team won 3rd place in the Novice Division at the 2023 Lime Rock Park Electrathon Competition

Background

We built an electric vehicle from scratch for an endurance competition, developing a car that could go the most distance in an hour.

Important Components

Two 12V Yellow Top Optima lead-acid batteries, providing a total voltage of 24V for high specific energy

Alltrax Motor Controller

DC Motor with 2800 RPM

Bicycle Disc Brakes

Tools Used

Fusion 360 & 3D Printer to design and print parts that protected the electrical components and connected awkward pieces such as the mirrors and fan

Soldering iron to connect electronics

Hot glue gun, string, and tape to connect parts to the chassis

Digital multimeter to measure voltage, current, and resistance at different parts of the circuit

Process

While originally developing solely the foundations of the vehicle including the chassis, mirrors, throttle, brakes, steering wheel, and chair, I ran into problems with the electrical system.

Seeing this challenge would eventually need to be overcome, I spearheaded the development of the circuit design. The given schematic for the motor controller appeared to be incorrect, so I researched, tested, and implemented my own adaptation inspired by the original to power the vehicle.

Circuit

The simplest loop in the circuit was connecting the controller pins to the potentiometer that controls the throttle as shown in the original schematic. The rest of the circuit was not as easy, and we could never produce power to rotate the motor sprocket following the original.

So, I continuously redesigned and tested the schematic until I found my final design, as shown above, that incorporated a brake light and display monitor. Once the vehicle was able to run, we focused on perfecting the other aspects.

Calculations

For the endurance aspect of the competition, we looked to maximize our performance by analyzing the weight and energy cost of the vehicle.

The competition required that the minimum weight of the driver with additional weight to the vehicle must be 80kg. Since both drivers were under 80kg, we stacked our motor, batteries, and all the electronics in the back and added our ballast to the front. This distribution proved advantageous, allowing the vehicle to better climb the hills in the course.

We stayed around 30-40mph throughout the entire race to conserve energy and maintain a relatively fast speed. We used more energy to climb hills and on the straight-aways than while turning or descending.

I measured the center-to-center distance between the motor and wheel to be about 16", ensuring little slack with the gear chain. Through online research, I determined a 6:1 motor-to-sprocket ratio to be most efficient for the endurance, turns, and hills of the competition.

Donovan Crowley

There's nothing a little duct tape and math can't solve

About Me

Projects

VTOL

ESP32, Raspberry Pi, Electromagnet, LiDAR, Computer Vision

Electric Vehicle

Circuit Design, Fusion360, Data Analysis

Web Scraper

API, BFS, A* Search, Machine Learning

Object Tracker

OpenCV, Object Detection, Hough Circle Transform, Computer Vision, Machine Learning

Photo Text Collage

OpenCV, K-Means, Machine Learning

Skills

Machine Learning

Comuter Vision

C++

Python

Java

HTML, CSS, JavaScript

Contact Me

Donovan Crowley

About Me

Projects

VTOL

ESP32, Raspberry Pi, Electromagnet, LiDAR, Computer Vision

Electric Vehicle

Circuit Design, Fusion360, Data Analysis

Web Scraper

API, BFS, A* Search, Machine Learning

Object Tracker

OpenCV, Object Detection, Hough Circle Transform, Computer Vision, Machine Learning

Photo Text Collage

OpenCV, K-Means, Machine Learning

Skills

Machine Learning

Comuter Vision

C++

Python

Java

HTML, CSS, JavaScript

Contact Me

Vertical Take-Off and Landing

Background

Process

Electric Vehicle

Background

Important Components

Tools Used

Process

Circuit

Calculations

Web Scraper

Object Tracker

Overview

SEARCH

TRAIN

TRACK

READ

Photo Text Collage