Doctor of Philosophy (Ph.D.)
Degree Granting Department
Engineering Computer Science
Sriram Chellappan, Ph.D.
Shaun Canavan, Ph.D.
Michael Coovert, Ph.D.
Nasir Ghani, Ph.D.
Neal Tempestt, Ph.D.
Driver Profiling, Intelligent Transportation Systems, Location Privacy, Supervised Machine Learning, Wearable Sensing
In this dissertation, we design algorithms to profile driver behavior from zero-permission sensors embedded in modern smartphones and wearables. These sensors are typically the accelerometer, gyroscope, magnetometer, pressure sensor and a few more than are now available in most modern smartphones and wearables. In order to profile driving behavior, we devised algorithms for detecting distraction while driving due to the use of modern-day smartphones (e.g., calling, texting and reading while driving) in real-time.
To do so, we conduct an experiment with 16 subjects on a realistic driving simulator, where each subject, where each subject carries a smartphone and a wearable worn on the right wrist.
Our simulator is the Computer Assisted Rehabilitation Environment system (CAREN) at the University of South Florida. As discussed later, the simulator is equipped with a realistic steering wheel, acceleration/ braking pedals, and a widescreen to visualize background vehicular traffic and is programmed to simulate multiple environmental conditions like daytime, nighttime, fog and rain/ snow. For the study, subjects were instructed to drive the simulator while performing a randomized sequence of activities that included texting, calling and reading from a phone while they were driving, during which the accelerometer and gyroscope in the smartphone and wearable were logging sensory data. We first started by exploring the feasibility of leveraging the accelerometer and gyroscope sensors in modern smartphones to detect instances of distracting driving activities. After achieving encouraging results, we explored the feasibility of using only the embedded accelerometer and gyroscope sensors of wrist wearable. The results were encouraging but gyroscope features were not of help. Finally, we used only accelerometer sensors from both smartphone and wrist wearable for profiling driver distraction due to phone usage. We used sensor fusion techniques and identified four intuitive features by fusing accelerometer readings from both smartphones and the wearable for classification. Features were identified such that the motion of the hand is tracked with respect to the phone during driving. Our performance evaluations reveal very good Precision, Recall, and F1-Scores for all the activities. Our system that fuses the accelerometer sensor readings from both smartphone and wearable produced the best results.
In this dissertation, we also explore location privacy breaches via processing accelerometer and gyroscope sensors from wrist wearables towards profiling subjects traveling in public transport systems. We design a technique to process the wearable accelerometer and gyroscope sensor data in order to identify routes taken by humans as they travel in public buses across the city. We successfully identified subject boarding/de-boarding a bus, left, right and U-turns taken by bus and starting/stopping of a bus. We feed activities detected along with timestamp, timetable and route map of the public bus as input to our algorithm for identifying route taken by the passenger when traveling on a public bus.
We believe that our study introduces new and potentially important applications and vulnerabilities of zero-permission embedded sensors for profiling in transportation systems.
Scholar Commons Citation
Goel, Bharti, "Algorithms To Profile Driver Behavior From Zero-permission Embedded Sensors" (2020). USF Tampa Graduate Theses and Dissertations.