Sylvia Bhattacharya | Georgia Southern University (original) (raw)

Uploads

Papers by Sylvia Bhattacharya

Transportation Engineering, 2022

SoutheastCon 2021, 2021

Roadway fatalities are increasing with a growing population and need for reliable transportation.... more Roadway fatalities are increasing with a growing population and need for reliable transportation. These fatalities can be mitigated by incorporating driver state information with current Driver Safety Systems (DSS). There are primarily two driver cognitive states: Focused and Distracted. These states can be predicted using Machine Learning algorithms (ML) such as Support Vector Machines (SVM), Adaptive Boosting (AdaBoost), and Artificial Neural Networks (ANN) using extracted biomedical features like Electroencephalography (EEG), Electromyography (EMG), Heart Rate and Eye Tracking. This literature review summarizes all biomedical signals that are used in the assessment of driver cognitive states. A thorough literature review in this field identifies eye tracking as the most efficient and quick technique of real time driver state identification. Hence, this paper outlines the latest techniques in eye tracking using oculometry. This review paper also highlights unique ocular feature extraction techniques that can be extremely useful for future researches conducted in the field of driver state recognition.

2020 SoutheastCon, 2020

Correlation studies between distracted driving and conversational tasks is investigated in this p... more Correlation studies between distracted driving and conversational tasks is investigated in this paper. It is a fact that as technology becomes more ubiquitous, we become more dependent on its integration into our daily routines, including driving tasks. With the integration of technology and cars, drivers can become easily distracted with the bombardment of secondary tasks such as cell phone usage, navigation systems, listening to podcasts, etc. We analyzed the cognitive correlations between distracted driving behavior based on electroencephalography (EEG) signals using conversation with co-passengers as a secondary driving task and a linear support vector machine (SVM) for classification. Results indicated that EEG features may provide a salient feature set for detecting driver distraction behavior. Furthermore, when compared with previous classification methods, the accuracy rating improved by 5% using SVM classification. Therefore, from these preliminary results, we posit that cognitive data can be used as a valid feature for classifying distracted driver behaviors. Future studies will further validate this result. The authors suggest, however, that a combination of EEG data with physiological features using SVM classification may provide a more robust system with improved performance accuracy.

2020 SoutheastCon, 2020

Automatic genre classification is a widely explored topic nowadays. It is an important task as th... more Automatic genre classification is a widely explored topic nowadays. It is an important task as the quantity of music released is sky-rocket. A survey says that alone Spotify releases about ten thousand songs a month. At the same time, users developed a taste of listening to music of different languages. Hence, it is very crucial to maintain a data base that can automatically classify songs of any language based on their genres. This project attempts to utilize a deep learning convolutional neural network technique to identify the genre of a song from twelve different languages all over the world. This project utilized a data base of sixty songs and mainly classified them into three genres: metal, hip hop, and pop. Each song is trimmed to be exactly five seconds long, and each five second "slice" is converted into a spectrogram, a visual representation of a song’s spectrum in the time domain. These spectrograms are divided into training and testing dataset. 75% of the data set is used for training and 25% is used for testing. A tensor flow deep learning convolutional neural network is used as an image classifier to automatically segregate the songs according to their genres. This deep learning model achieved a classification accuracy of 93.3% for all of the twelve languages considered in this database.

College Teaching, 2021

In the Spring term of 2020, nearly 90% of higher education institutions in the United States were... more In the Spring term of 2020, nearly 90% of higher education institutions in the United States were forced to transition from primarily face-to-face (F2F) instruction to various modes of remote or online instruction in response to the COVID-19 pandemic. State-funded colleges and universities in Georgia were mandated to do the same in April of 2020, which led to a system-wide hiatus in face-to-face instruction while instructors prepared to return to all-remote teaching. This study examined the effects of this transition to Emergency Remote Instruction (ERI) at six institutions in Georgia, using a survey completed by 910 instructors who made that transition in at least one course in the Spring term of 2020. 65% of the instructors taught remotely or online for the first time after the transition. Instructors reported accessing a variety of institutional, collegial, and internet resources to aid in the transition, leading 53.4% of them to express that they were adequately prepared for ERI. Once classes resumed online, instructors found themselves to be needing much more time for remote instruction than their previous F2F instruction. From a one-word summary description of their experience, instructors reported that it led them to be challenged, stressed, overwhelmed, and exhausted. [ABSTRACT FROM AUTHOR] Copyright of College Teaching is the property of Taylor & Francis Ltd and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

SoutheastCon 2017, 2017

Electroencephalography is widely used to record neural activity with electrodes positioned at spe... more Electroencephalography is widely used to record neural activity with electrodes positioned at specific locations on a human scalp. These recorded signals are interfaced with a computer which is referred to as noninvasive Brain Computer Interface (BCI). An important application of this technology is to help facilitate the lives of the tetraplegic through assimilating human brain impulses and converting them into mechanical motion. However, BCI systems are remarkably challenging to implement as recorded brain signals can be unreliable and vary in pattern throughout time. In this paper, a novel classifier structure is proposed to classify different types of imaginary motions (left hand, right hand, and imagination of words starting with the same letter) across multiple sessions using an optimized set of electrodes for each user. The proposed technique uses raw brain signals obtained utilizing 32 electrodes and classifies the imaginary motions using Artificial Neural Networks (ANN). To enhance the classification rate and optimize the set of electrodes of each subject, a majority voting system combining a set of simple ANNs is used. This electrode optimization technique achieved classification accuracies of 69.83%, 94.04% and 84.56% respectively for the three subjects considered in this study.

SoutheastCon 2018, 2018

1- Video capture methods are useful for identifying information about a driver's visual behav... more 1- Video capture methods are useful for identifying information about a driver's visual behavior and their intent to perform a driving maneuver. They can capture data on where drivers are looking during primary and secondary driving tasks, which can be important for identifying driver intentions and roadway distractions. In this study the eye-glance and head movement of a driver performing a lane-merge task in a driving simulator is observed. Drivers were given a multitask driving scenario while performing lane-change and highway merging maneuvers. Multiple raters were used to assess the reliability of using eye-glance and head motion data for analyzing driver performance. The inter-rater reliability method was used to obtain a driver profile. The purpose of this study is to provide an analytical framework for assessing an elderly (65+) driver's performance during lane-change and highway merging multitask maneuvers using the video capture method. This work contributes to the discovery of knowledge and deeper insight about how a driver's behavior and physical state may directly influence their driving performance and profile.

2021 IEEE 94th Vehicular Technology Conference (VTC2021-Fall), 2021

Autonomous Driving has been continuously improving with the integration of bio-physiological data... more Autonomous Driving has been continuously improving with the integration of bio-physiological data and deep learning systems. Passenger-driver behavior from psychological processes can provide insight into the trust levels of the dyad dynamic during an open-road drive. Measuring passenger trust can benefit in speeding up the adoption of self-driving cars. This area has not been fully explored from the passenger's point of view and can yield results that could potentially propel the progress of driver safety systems in autonomous driving. A novel approach to correlating the passenger-driver trust levels can be elicited from brain-based indicators that are extracted from time-locked electroencephalogram (EEG) signals that capture both sensory and cognitive processes. In this paper, we propose a trust identification technique utilizing evoked response potential (ERP) events such as P300 along with beta/alpha frequency ratios from specific passengers during braking, aggressive acceleration, and lane changing scenarios. The results obtained from the decomposition of these EEG signals into the frequency bands and application of machine learning techniques on data collected from frontal and parietal electrodes during these driving events prove the feasibility of this study. By examining additional passenger-driver pairs with varying social interaction and trust profiles in the future, we can strengthen the existence of a cognitive correlation between passenger-driver behavior and thus improve the efficiency of driver safety systems.

2019 SoutheastCon, 2019

Using a driver simulator to detect driver behavior is a very common but challenging field that st... more Using a driver simulator to detect driver behavior is a very common but challenging field that started developing rapidly since after World War II. This technique is called virtual reality simulation. Virtual reality simulation is very important in case of driving as it provides a very safe environment for conducting research. In this technique, a virtual world is developed which replicates the real world and helps researchers to collect data from subjects to understand various human driving behavior. In this paper, lateral variation characteristics (speed variance and lane positioning) are monitored on a virtual highway, to help understand the cause of innumerous accidents on US highways due to failure of lane maintenance. Data was collected from five healthy subjects in two sessions where each driver had to perform two different tasks. The result implies that any secondary task along with driving distracts the driver to some extent and involves risk on the road. This study will quantify the extent to which secondary driving tasks impact primary task. The mental workload of the drivers are also determined in order to understand the mental demand of driving on the simulator.

2021 IEEE 94th Vehicular Technology Conference (VTC2021-Fall), 2021

2020 SoutheastCon, 2020

Engineering is a field where acquiring practical knowledge is very important along with theoretic... more Engineering is a field where acquiring practical knowledge is very important along with theoretical studies. Electronics I is a subject where an engineering student gets introduced to building circuits for the first time. Building a circuit on a bread board from a circuit diagram is challenging initially for most students. In this paper, a cognitive behavioral study is done to estimate a better way of teaching this lab. In this study, pilot data is collected from 22 students over a semester to understand their progress in handling the operational process of various electronics equipment. A comparative analysis is performed to see if the students' pace of learning is more when assigned in a group versus individual study. Before each lab, two theoretical classes on Electronics I was conducted so that students learn the theory before implementing it in hands-on experiments in lab class. A survey was conducted during mid-semester and at the end of the semester to compare the comprehension level of each student individually and cooperatively at critical time points throughout the class. Results shows that more than 90% of students agreed that working with two people in a lab group gives the most optimized output. Learning is much quicker and more in-depth when a hands-on project is performed in a group. It is also examined how a theoretical class helps in performing laboratory tasks. This study helps understand a better way to conduct a lab class in electrical engineering so that students gain interest in the subject and at the same time learn effectively.

Video capturing technique is useful for detecting driver's visual behavior and their intent t... more Video capturing technique is useful for detecting driver's visual behavior and their intent to perform a driving task. Video analysis helps in understanding driver intentions and roadway distractions by looking at the driver's eye positions while s/he is performing his primary and/or secondary driving task. Conducting research on a driver simulator is preferred by many researchers as it provides a safe environment to analyze various roadway conditions. In this study, driver simulator is used to collect data from ten subjects (7 males and 3 females) to determine whether eye glance behavior is affected by secondary tasks while driving. Subjects are instructed to drive in a highway scenario with three co-passengers in the car and also alone, while changing lanes whenever possible. Eye glance of the driver is observed for 90 seconds for each scenario and the compared result implies that driving with co passengers increased eye glance rapidly. An overall average frequency of 0.11...

SoutheastCon 2016, 2016

Using Electroencephalography (EEG) to detect imaginary motions from brain waves to interface huma... more Using Electroencephalography (EEG) to detect imaginary motions from brain waves to interface human and computer is a very nascent and challenging field that started developing rapidly in the past few decades. This technique is termed as Brain Computer Interface (BCI). BCI is extremely important in case of people who are incapable of communicating due to spinal cord injury. This technique uses the brain signals to make decisions, control and communicate with the world using brain integration with peripheral devices and systems. In this paper, in order to classify imaginary motions, raw data are used to train a system of neural networks with a majority vote output. EEG data for 3 subjects are used from the BCI Competition III dataset V. Each subject has data collected in three sessions representing three different types of imaginary motions. Using an optimized set of electrodes, classification accuracy was optimized for the three users as a group. A cross validation method is applied to improve the reliability of the generated results. The optimization resulted in an electrode structure consisting of 15 electrodes with a relatively high classification accuracy of almost 80%.