AutoBleedr: Automating Hydraulic Brake Bleeding in High Performance Bikes; Analyzing the 2018 Uber Tempe crash using User Configuration 77 views

My computer engineering technical project and STS research project are connected through a shared inclusion of automation. Automation involves having mechanical and/or computation systems perform tasks usually done by humans, usually to save time or reduce human effort. My technical project used automation because it attempted to automate the brake bleeding process of mountain bikes, and my STS research project examined the Uber automated car crash in Tempe, Arizona and how mismatched user configurations led to it. Though both deal with automation differently, the theme of replacing manual processes with automated systems is prevalent in both. My technical project was to develop a device that automated enough of the mountain bike brake bleeding process so that the user would only have to deal with set up and clean up. Brake bleeding is a messy process that requires precision and automating it saves time for both recreational bikers and mechanics. To this end, my capstone team designed and integrated several mechanical components and sensors with a printed circuit board (PCB) that managed activating mechanical components based on a timer and sensors’ outputs. The mechanical components meant that users wouldn’t have to interact with the bike brake, while the sensors meant the user didn’t have to adjust the PCB once started. By the end of the semester, we completed a prototype that could handle bleeding a mock bike brake, and one of my teammates hopes to finalize the design and market it from their startup. My STS research project also deals with automation by examining the Tempe, Arizona Uber crash in 2018 that resulted in the death of a pedestrian. The Uber car involved in the crash was under testing, being controlled by an experimental automated driving system (ADS) with an overseeing backup driver. I analyzed this incident through the lens of user configuration, developed by Steve Woolgar, examining how designers embedded their user assumptions into their design choices of the ADS and the training of the driver. In doing so, I found that the configured user was greatly mismatched with the actual driver, especially in the domain of attention, significantly contributing to the accident. The results of my analysis served as a reminder to engineers of how crucial making correct assumptions about the end user is for preventing accidents involving automation. Working on my STS research project at the same time as my technical project added great value to the development of the automated brake bleeder. My STS project demonstrated that even as automation reduced human input, engineers still needed to consider the behavior of users. This insight influenced the design of my technical project as I emphasized usability over complexity by minimizing the number of buttons and designing the device to stop automatically rather than relying on human judgment. In short, working on my STS project made me realize that automation isn’t only about minimizing human effort but also about considering human behavior, a principle I simultaneously applied to my technical project.

BS (Bachelor of Science)

Automation; Brakes; Uber; User-Configuration; Cars

School of Engineering and Applied Science Bachelor of Science in Computer Science and Computer Engineering Technical Advisor: Caroline Crockett STS Advisor: Benjamin Laugelli Technical Team Members: An Huynh, Landon Campbell, Paul Wiskow, Thomas Keyes

English

All rights reserved by the author (no additional license for public reuse)

2026-05-08