Projects
Currently, I am involved with several projects which are directly or indirectly related to Smarthome activity recognition. I worked on several other projects previously related to distinguished field of Computer Science.
Current Projects
[1] Elderly Care: A Context Aware Smart Computing Model for Predicting Cognitive Health In this project, we investigate the challenges of improving the automatic assessment of dementia, by better exploiting the emerging physiological sensors in conjunction with ambient sensors in a real field environment with IRB approval. We hypothesize that the cognitive health of older individuals can be estimated by tracking their daily activities and mental arousal states. We employ signal processing on wearable sensor data streams (e.g., Electrodermal Activity (EDA), Photoplethysmogram (PPG), accelerometer (ACC)) and machine learning algorithms to assess cognitive impairments and its correlation with functional health decline.
Publications: IEEE Conference on Connected Health: Application, Systems and Engineering Technology (CHASE 2016), International Symposium on Wearable Computers (ISWC) (Under Review)
Collaborators: Nirmalya Roy (UMBC), Elizabeth Galik (University of Maryland Baltimore), Aryya Gangopadhya (UMBC)
[2] Smarthome Activity Recognition: In this project we are exploiting the spatiotemporal constraints and correlations for recognizing intra- and inter-relationships across multiple inhabitants at the micro, meso and macro-level of activities in smarthome environment.
Publications: 36th IEEE International Conference on Distributed Computing Systems 2016 (ICDCS 2016)
Collaborators: Nirmalya Roy (UMBC), Archan Misra (Singapore Management University)
[3] Mobeacon: iBeacon Assisted Smarthome Activity Recognition Framework A mobile phone and iBeacon sensorbased smart home activity recognition system. We investigated the viability of extending Bagging Ensemble Learning (BEL) and Packaged Naive Bayes (PNB) classi cation algorithms for high-level activity recognition on smartphone. We incorporated the semantic knowledge of the testing environment and used that with the built-in adaptive learning models on smartphone to ease the ground truth data annotation.
[4] GeSmart: Smart Jewelry Based Gestural Activity Recognition System: In this project we are working on an energy efficient wearable smart earring based Gestural Activity recognition model for detecting a combination of speech and non-speech events. To capture the Gestural Activities, we are using only the accelerometer sensor inside our smart earring due to its energy efficient operations and ubiquitous presence in everyday wearable devices.
[5] Light-Weight Implementation of Dynamic Bayesian Network framework in Wearable smart devices: In this project we are working on an efficient algorithm in terms of accuracy and computational overhead of Dynamic Bayesian Network especially focusing on Hierarchical Dynamic Bayesian Network and it's variants in mobile system. We will employ our light-weight algorithm in real-time Smarthome activity recognition system to evaluate accuracy and computational overhead efficiency. We implemented our proposed Hierarchical Dynamic Bayesian Network in Android platform using JAHMM api. We are extending it to the next level of light-weight computations.
Previous Projects
Elderly Adult Safety Smartphone based approach: We have developed an algorithmic model, monitored and documented elderly people`s daily activities by using the gyroscope and accelerometer of a smartphone and with the use of those data and model, we calculated how much activity is required or overdone for a subject in order to maintain a healthy lifestyle. More importantly, we built a real time system that could not only judge what basic activity the subject is currently doing, but also protect the subject from possible injury that might happen to the subject if abnormal data is received.
Trustworthy Computing in Pervasive Environment: For privacy-aware context-based applications in pervasive environment, we work on a formal collaborative model that preserves users’ privacy without involving a trusted mediator between the requester and the service provider. In addition, a unique privacy measuring technique is introduced to provide a fundamental basis for quantifying the privacy of the requester at the time of requesting the service. We evaluated and proved that our technique can provide better privacy in pervasive environment.
Considerations in Designing Human-Computer Interfaces for Elderly People: We explore common problems the elderly face when using computing devices and solutions developed for these problems. Difficulties ultimately fall into several categories: cognition, auditory, haptic, visual, and motor-based troubles. We also present an idea for a new adaptive operating system with advanced customizations that would simplify computing for older users.
A Social & Technology Support Program for Veteran Mental Health: We developed a model for both peer mentor and veterans to acquire, monitor and alert activities in case of significant change from norm. In this regards, along with the a group of psychologists from Medical College of Wisconsin, we settled a set of questionnaire and a hand held device (Android) based ecological momentary assessment survey system for risky veterans. We also developed a web based dashboard and control system for weekly survey questions. We are collaborating with Dryhootch, a veterans social support organization, Medical College of Wisconsin, University of Wisconsin at Milwaukee for this pilot project.
Improving Energy Efficiency of GPU based General-Purpose Scientific Computing: This is a class project for the course Parallel and Distributed System at Marquette University with Dr. Rong Ge. Here, we describe our technique for a coordinated measurement approach that combines real total power measurement and per-component power estimation. This work also analyses energy consumption of MAGMA (Matrix Algebra on GPU and Multicore Architectures). Based on our experiments of dynamically estimating the runtime GPU power consumption, through preliminary empirical validation of a number of GPGPU benchmarks, we demonstrated that our framework can be robustly used to measurably improve the energy efficiency of various GPGPU programs.
A Mobile Application for Creating Flashcards and Social StoriesT for Children with Autism: The purpose of this research study is to evaluate a mobile app – ‘iCanLearn’ in teaching social skills or daily activities (like how to brush your teeth, how to get ready for school, etc.) to a child with autism spectrum disorder (ASD). iCanLearn is a flashcard based learning app developed in Marquette University Ubicomp Lab available for free download on iPhones and iPads. This app lets users create their own flashcards with text, pictures and audio features. They can use these features to teach a certain activity in a step-by-step manner. This evaluation will be done by parents or primary caregivers of a child with ASD. We would like to see if the app is easy enough to use and can be used in different situations as needed.