Senior Research Project Proposal
Name: Sarah Wooders
BASIS Advisor: Porter McDonald
Title
BASIS Advisor: Porter McDonald
Title
Safer Teen
Driving in the Age of Smartphones
Background
Throughout high school I
have had two main interests – Economics, and Computer Science. From 8th
to 9th grade, I took and audited Economics classes at the University
of Arizona, including Intermediate Microeconomics and Experimental
Economics. In 11th grade, I
was an avid competitive programmer, which helped me become familiar with many
common algorithms and programming techniques as well as helping me become an
extremely fast coder. As a result of my dual background in Economics and
Computer Science, last summer, I interned with MobLab, a Pasadena based startup
which made educational games utilized in Economics, Game Theory, Psychology,
and Political Science classes. There as an intern, my project was to design and
implement computer players, and I was fascinated by how algorithms could be
used to emulate and understand behaviors studied in Economics.
A major part of the
challenge of creating computer players was making them capable of learning. If
a human played in a particular way, I wanted my computer player to be able to
recognize that and take advantage of it. As a result, a large portion of my
project was dedicated to writing algorithms to learn from historical data. This
led me to become very interested in algorithms and statistical models that
learned from data.
For my SRP, I wanted to
work as an intern for a company which used interesting algorithms in software
development. I found the perfect company here in Tucson, Metropia, which has
developed an app that incentivizes drivers to cooperate in order to reduce
traffic congestion. The core of the app predicts congestion levels based off
user data, and then rewards users with gift cards for adjusting their driving
(for example, leaving before rush hour) in a manner that decreases congestion.
Currently, Metropia is beginning to work on developing a feature for their app
which assesses driving quality based off the motions of a smartphone in a car. Vehicle motion can be examined by looking at
the time series data from accelerometers and the gyroscope in phones located
inside the moving car. Thus, in order to identify different motions, different
statistical and machine learning models must be used.
Statement of Purpose
As
a research intern for Metropia, I will
be partaking in the development of the driving assessment feature of Metropia’s
app, responsible for aspects of design, development, and testing. I hope to
develop a feature which is accurate and efficiently able to driving quality by using smartphone accelerometer and
gyroscope data to look at sudden breaks, sharp turns, or speeding, which could
be used to decrease insurance premiums, give feedback for new drivers, and
otherwise incentivize safer driving. I hope to have everything in place for the feature
by April, ready for launch this summer.
Methodology
I will be working
as part of the Metropia’s R&D team under the supervision of Xian-Bian Hu to
develop algorithms to assess driving quality based off accelerometer and
gyroscope data. I will also be working on the Metropia’s Teen Driver program
under the supervision of Chris Coleman.
0 comments:
Post a Comment