Date of Award

Summer 2016

Document Type


Degree Name

Master of Science (MS)


Computer Science

Committee Director

Tamer Nadeem (Director)

Committee Member

Michele Weigle

Committee Member

Stephan Olariu


We introduce a solution that uses the availability of heavy crowds and their smart devices, to gain more result as to where potential parking is possible. By leveraging the raw magnetometer, gyroscope, and accelerometer data, we are able to detect parking spots through the natural movement exerted by the walking pedestrians on the sidewalks beside the streets. Dating back as far as 2013, a very large portion of pedestrians composing the crowds on the sidewalk, possessed at least one smart device in their hand or pocket14]. It is this statistic that fuels our application, in which we depend on crowds or even a steady rate of pedestrians, telling others around them where unoccupied parking sport are, without making a single bit of noise. In other words, we use the walking pedestrians’ cellphone sensors to classify the sidewalk parking spots as occupied and vacant. The more pedestrians walking on the sidewalk, the more accurate our application works. As the years and technological advances both increase, we predict that the number of smart devices will only increase, allowing our solution to become much more precise and useful.

The biggest contribution of our study can be summarized as follows:

• Implementation of Magnopark; a high accuracy parking spot localization system using internal smart phone sensors

• Evaluation and test of Magnopark in different situations and places

• Test of Magnopark for different users with different walking habits and speed

• Development of an algorithm to detect the users’ stride, speed, and direction change

• Building a classification model based on the features extracted from the cellphone sensors

• Pushing the classified data to the cloud for the drivers’ use