Computer Engineering Commons^™

Analysis And Control Of Fiber-Reinforced Elastomeric Enclosures (Frees), Soheil Habibian

Master’s Theses

While rigid robots are extensively used in various applications, they are limited in the tasks they can perform and can be unsafe in close human-robot interactions. Soft robots on the other hand surpass the capabilities of rigid robots in several ways, such as compatibility with the work environments, degrees of freedom, manufacturing costs, and safe interactions with the environment. This thesis studies the behavior of Fiber-Reinforced Elastomeric Enclosures (FREEs) as a particular type of soft pneumatic actuator that can be used in soft manipulators. A dynamic lumped-parameter model is created to simulate the motion of a single FREE under various …

Dedicated Hardware For Machine/Deep Learning: Domain Specific Architectures, Angel Izael Solis

Open Access Theses & Dissertations

Artificial intelligence has come a very long way from being a mere spectacle on the silver screen in the 1920s [Hml18]. As artificial intelligence continues to evolve, and we begin to develop more sophisticated Artificial Neural Networks, the need for specialized and more efficient machines (less computational strain while maintaining the same performance results) becomes increasingly evident. Though these “new” techniques, such as Multilayer Perceptron’s, Convolutional Neural Networks and Recurrent Neural Networks, may seem as if they are on the cutting edge of technology, many of these ideas are over 60 years old! However, many of these earlier models, at …

An Explainable Autoencoder For Collaborative Filtering Recommendation, Pegah Sagheb Haghighi, Olurotimi Seton, Olfa Nasraoui

Faculty and Staff Scholarship

Autoencoders are a common building block of Deep Learning architectures, where they are mainly used for representation learning. They have also been successfully used in Collaborative Filtering (CF) recommender systems to predict missing ratings. Unfortunately, like all black box machine learning models, they are unable to explain their outputs. Hence, while predictions from an Autoencoderbased recommender system might be accurate, it might not be clear to the user why a recommendation was generated. In this work, we design an explainable recommendation system using an Autoencoder model whose predictions can be explained using the neighborhood based explanation style. Our preliminary work …

Emergence Of Addictive Behaviors In Reinforcement Learning Agents, Vahid Behzadan, Roman Yampolskiy, Arslan Munir

Faculty and Staff Scholarship

This paper presents a novel approach to the technical analysis of wireheading in intelligent agents. Inspired by the natural analogues of wireheading and their prevalent manifestations, we propose the modeling of such phenomenon in Reinforcement Learning (RL) agents as psychological disorders. In a preliminary step towards evaluating this proposal, we study the feasibility and dynamics of emergent addictive policies in Q-learning agents in the tractable environment of the game of Snake. We consider a slightly modified version of this game, in which the environment provides a “drug” seed alongside the original “healthy” seed for the consumption of the snake. We …

Instance Segmentation And Object Detection In Road Scenes Using Inverse Perspective Mapping Of 3d Point Clouds And 2d Images, Chungyup Lee

Electronic Theses and Dissertations

The instance segmentation and object detection are important tasks in smart car applications. Recently, a variety of neural network-based approaches have been proposed. One of the challenges is that there are various scales of objects in a scene, and it requires the neural network to have a large receptive field to deal with the scale variations. In other words, the neural network must have deep architectures which slow down computation. In smart car applications, the accuracy of detection and segmentation of vehicle and pedestrian is hugely critical. Besides, 2D images do not have distance information but enough visual appearance. On …

Smart Parking Deck, Ryne Turner, Matthew Mcdade, Julie Aichinger, Laveréna Wienclaw

Williams Honors College, Honors Research Projects

The Smart Parking Deck employs elementary circuit design elements and mobile application development. Each device module uses laser proximity sensors to check the availability of an individual parking space and a Zigbee unit to communicate with the adjacent device module. The modules are connected to a network hub that manages all of the incoming and outgoing parking data. This data is displayed on the mobile application. The system is easily manageable and energy efficient, significantly decreasing the costs associated with other smart parking systems on the market. This system is aimed at decreasing commute time for students by allowing them …

Islands Of Fitness Compact Genetic Algorithm For Rapid In-Flight Control Learning In A Flapping-Wing Micro Air Vehicle: A Search Space Reduction Approach, Kayleigh E. Duncan

Browse all Theses and Dissertations

On-going effective control of insect-scale Flapping-Wing Micro Air Vehicles could be significantly advantaged by active in-flight control adaptation. Previous work demonstrated that in simulated vehicles with wing membrane damage, in-flight recovery of effective vehicle attitude and vehicle position control precision via use of an in-flight adaptive learning oscillator was possible. Most recent approaches to this problem employ an island-of-fitness compact genetic algorithm (ICGA) for oscillator learning. The work presented provides the details of a domain specific search space reduction approach implemented with existing ICGA and its effect on the in-flight learning time. Further, it will be demonstrated that the proposed …

Towards Data And Model Confidentiality In Outsourced Machine Learning, Sagar Sharma

Browse all Theses and Dissertations

With massive data collections and needs for building powerful predictive models, data owners may choose to outsource storage and expensive machine learning computations to public cloud providers (Cloud). Data owners may choose cloud outsourcing due to the lack of in-house storage and computation resources or the expertise of building models. Similarly, users, who subscribe to specialized services such as movie streaming and social networking, voluntarily upload their data to the service providers' site for storage, analytics, and better services. The service provider, in turn, may also choose to benefit from ubiquitous cloud computing. However, outsourcing to a public cloud provider …

Abusive And Hate Speech Tweets Detection With Text Generation, Abhishek Nalamothu

Browse all Theses and Dissertations

According to a Pew Research study, 41% of Americans have personally experienced online harassment and two-thirds of Americans have witnessed harassment in 2017. Hence, online harassment detection is vital for securing and sustaining the popularity and viability of online social networks. Machine learning techniques play a crucial role in automatic harassment detection. One of the challenges of using supervised approaches is training data imbalance. Existing text generation techniques can help augment the training data, but they are still inadequate and ineffective. This research explores the role of domain-specific knowledge to complement the limited training data available for training a text …

Llvm-Ir Based Decompilation, Ilsoo Jeon

Browse all Theses and Dissertations

Decompilation is a process of transforming an executable program into a source-like high-level language code, which plays an important role in malware analysis, and vulnerability detection. In this thesis, we design and implement the middle end of a decompiler framework, focusing on Low Level Language properties reduction using the optimization techniques, propagation and elimination. An open-source software tool, dagger, is used to translate binary code to LLVM (Low Level Virtual Machine) Intermediate Representation code. We perform data flow analysis and control flow analysis on the LLVM format code to generate high-level code using a Functional Programming Langauge (FPL), Haskell. The …

Internet Of Things In Smart Agriculture: Enabling Technologies, Abdul Salam, Syed Shah

Faculty Publications

In this paper, an IoT technology research and innovation roadmap for the field of precision agriculture (PA) is presented. Many recent practical trends and the challenges have been highlighted. Some important objectives for integrated technology research and education in precision agriculture are described. Effective IoT based communications and sensing approaches to mitigate challenges in the area of precision agriculture are presented.

Multirobot Confidence And Behavior Modeling: An Evaluation Of Telerobotic Performance And Efficiency, Nathan Lucas

Wayne State University Dissertations

There is considerable interest in multirobot systems capable of performing spatially distributed, hazardous, and complex tasks as a team. There is also growing interest in manned-unmanned teams leveraging the unique abilities of humans and automated machines working alongside each other. The limitations of human perception and cognition affect the ability of operators to integrate information from multiple mobile robots, switch between their spatial frames of reference, and divide attention among many sensory inputs and command outputs. Automation is necessary to help the operator manage increasing demands as the number of robots scales up. However, more automation does not necessarily equate …

Automatic Resource Management And Performance Optimization In Clusters, Yudi Wei

Wayne State University Dissertations

Virtual machine is a primary way to increase resource utilizations in data centers by encapsulating multi-resource demands for applications and providing performance isolation. Moreover, the resource configuration can change on the fly to satisfy performance target. Container is another popular way for fine-grained multi-resource allocation. In this disser- tation work, we aim to design and implement an automatic resource management system to improve application performance, optimize system efficiency and job completion times in virtual and physical clusters respectively.

For large-scale applications hosted in data center, automatic resource configuration is crucial to service availability and quality. The workload dynamics, cloud dynamics …

Mediaeval2019: Flood Detection In Time Sequence Satellite Images, Palavi Jain, Bianca Schoen-Phelan, Robert J. Ross

Conference papers

In this work, we present a flood detection technique from time series satellite images for the City-centered satellite sequences (CCSS) task in the MediaEval 2019 competition [1]. This work utilises a three channel feature indexing technique [13] along with a VGG16 pretrained model for automatic detection of floods. We also compared our result with RGB images and a modified NDWI technique by Mishra et al, 2015 [15]. The result shows that the three channel feature indexing technique performed the best with VGG16 and is a promising approach to detect floods from time series satellite images.

Autonomous Combat Robot, Andrew J. Szabo Ii, Chris Heldman, Tristin Weber, Tanya Tebcherani, Holden Leblanc, Fabian Ardeljan

Williams Honors College, Honors Research Projects

This honors project will also serve as an engineering senior design project.

The objective is to design and build the software and electrical systems for a 60 lb weight class combat robot that will function autonomously and outperform manually driven robots during competition.

While running autonomously, the robot will use LiDAR sensors to detect and attack opponent robots. This robot will also be able to be remote controlled in manual mode. This will mitigate the risk in case the autonomy or sensors fail. LED lights on the robot will indicate whether it is in autonomous or manual mode. The system …

Safe Pass, Alycia Riese, Julia Hariharan, Greg Synek, Jonathan Hall

Williams Honors College, Honors Research Projects

The purpose of this project is to design a sensor to be mounted on Class IV and higher vehicles to detect on-coming traffic. If traffic has been detected, the system is to warn drivers behind the stopped vehicle that passing is unsafe. The vehicle detection is to be implemented using a LiDAR detection method along with signal processing. A wireless transceiver is to transmit from the front radar module to the rear warning indicator module when the conditions are unsafe for passing. The project goals are to increase road safety and maintain traffic flow. The report details the challenges due …

Speech Enabled Navigation In Virtual Environments, Raksha Rajashekar

Browse all Theses and Dissertations

Navigating in a Virtual Environment with traditional input devices such as mouse, joysticks and keyboards provide limited maneuverability and is also time consuming. While working in a virtual environment, changing parameters to obtain the desired visualization requires time to achieve by manually entering parameter values in an algorithm to test outcomes. The following thesis presents an alternate user interface to reduce user efforts, while navigating within the Virtual Environment. The user interface is an Android application which is designed to accommodate spoken commands. This Speech Enabled User Interface termed as the Speech Navigation Application (SNA), provides the user with an …

Kbot: Knowledge-Enabled Personalized Chatbot For Self-Management Of Asthma In Pediatric Population, Dipesh Kadariya

Browse all Theses and Dissertations

Asthma, chronic pulmonary disease, is one of the major health issues in the United States. Given its chronic nature, the demand for continuous monitoring of patient’s adherence to the medication care plan, assessment of their environment triggers, and management of asthma control level can be challenging in traditional clinical settings and taxing on clinical professionals. A shift from a reactive to a proactive asthma care can improve health outcomes and reduce expenses. On the technology spectrum, smart conversational systems and Internet-of-Things (IoTs) are rapidly gaining popularity in the healthcare industry. By leveraging such technological prevalence, it is feasible to design …

Conditional Dilated Attention Tracking Model - C-Datm, Tyler Clayton Highlander

Browse all Theses and Dissertations

Current commercial tracking systems do not process images fast enough to perform target-tracking in real- time. State-of-the-art methods use entire scenes to locate objects frame-by-frame and are commonly computationally expensive because they use image convolutions. Alternatively, attention mechanisms track more efficiently by mimicking human optical cognitive interaction to only process small portions of an image. Thus, in this work we use an attention-based approach to create a model called C-DATM (Conditional Dilated Attention tracking Model) that learns to compare target features in a sequence of image-frames using dilated convolutions. The C-DATM is tested using the Modified National Institute of Standards …

Leveraging Schema Information For Improved Knowledge Graph Navigation, Rama Someswar Chittella

Browse all Theses and Dissertations

Over the years, the semantic web has emerged as a new generation of the world wide web featuring advanced technologies and research contributions. It has revolutionized the usage of information by allowing users to capture and publish machine-understandable data and expedite methods such as ontologies to perform the same. These ontologies help in the formal representation of a specified domain and foster comprehensive machine understanding. Although, the engineering of ontologies and usage of logic have been an integral part of the web semantics, new areas of research such as the semantic web search, linking and usage of open data on …

Static Evaluation Of Type Inference And Propagation On Global Variables With Varying Context, Ivan Frasure

Browse all Theses and Dissertations

Software reverse engineering (SRE) is a broad field with motivations ranging from verifying or documenting gordian source code files to understanding and reimplementing binary object files and executables. SRE of binaries is exceptionally compelling and challenging due to large amounts of information that can be lost in the compilation progress. A central area in SRE is type inference. Type inference is built around a fundamental step in understanding the behavior of a binary, recovering the types of data in the program. Type inference has many unique techniques in both static and dynamic type inference systems that have been implemented in …

Anticipation In Dynamic Environments: Deciding What To Monitor, Zohreh A. Dannenhauer

Browse all Theses and Dissertations

In dynamic environments, external changes may occur that may affect planning decisions and goal choices. We claim that an intelligent agent should actively watch for what can go wrong and anticipate changes in the environment that allows the changing of its plan or changing of a given goal. In this thesis, we focus on the relationship between perception, act, interpretation, and planning. We claim that these components are not independent and need to interact with each other to help the agent succeed in achieving its goals and plans. If newly encountered world information affects the plan, the agent adapts to …

Software Implementations And Applications Of Elliptic Curve Cryptography, Kirill Kultinov

Browse all Theses and Dissertations

Elliptic Curve Cryptography (ECC) is a public-key cryptography system. Elliptic Curve Cryptography (ECC) can achieve the same level of security as the public-key cryptography system, RSA, with a much smaller key size. It is a promising public key cryptography system with regard to time efficiency and resource utilization. This thesis focuses on the software implementations of ECC over finite field GF(p) with two distinct implementations of the Big Integer classes using character arrays, and bit sets in C++ programming language. Our implementation works on the ECC curves of the form y^2 = x^3 + ax + b (mod p). The …

Rules With Right Hand Existential Or Disjunction With Rowltab, Sri Jitendra Satpathy

Browse all Theses and Dissertations

One hotly debated research topic is, “What is the best approach for modeling ontologies?”. In the earlier stages of modeling ontologies, researchers have favored the usage of description logic to capture knowledge. One such choice is the Web Ontology Language (OWL) that is based on description logic. Many tools were designed around this principle and are still widely being used to model and explore ontologies. However, not all users find description logic to be intuitive, at least not without an extensive background in formal logics. Due to this, researchers have tried to explore other ways that will enable such users …

Recognition Of Incomplete Objects Based On Synthesis Of Views Using A Geometric Based Local-Global Graphs, Michael Christopher Robbeloth

Browse all Theses and Dissertations

The recognition of single objects is an old research field with many techniques and robust results. The probabilistic recognition of incomplete objects, however, remains an active field with challenging issues associated to shadows, illumination and other visual characteristics. With object incompleteness, we mean missing parts of a known object and not low-resolution images of that object. The employment of various single machine-learning methodologies for accurate classification of the incomplete objects did not provide a robust answer to the challenging problem. In this dissertation, we present a suite of high-level, model-based computer vision techniques encompassing both geometric and machine learning approaches …

Knowledge Graph Reasoning Over Unseen Rdf Data, Bhargavacharan Reddy Kaithi

Browse all Theses and Dissertations

In recent years, the research in deep learning and knowledge engineering has made a wide impact on the data and knowledge representations. The research in knowledge engineering has frequently focused on modeling the high level human cognitive abilities, such as reasoning, making inferences, and validation. Semantic Web Technologies and Deep Learning have an interest in creating intelligent artifacts. Deep learning is a set of machine learning algorithms that attempt to model data representations through many layers of non-linear transformations. Deep learning is in- creasingly employed to analyze various knowledge representations mentioned in Semantic Web and provides better results for Semantic …

Leveraging Blockchain To Mitigate The Risk Of Counterfeit Microelectronics In Its Supply Chain, Aman Ali Pogaku

Browse all Theses and Dissertations

System on Chip (SoC) is the backbone component of the electronics industry nowadays. ASIC and FPGA-based SoCs are the two most popular methods of manufacturing SoCs. However, both ASIC and FPGA industries are plagued with risks of counterfeits due to the limitations in Security, Accountability, Complexity, and Governance of their supply chain management. As a result, the current practices of these microelectronics supply chain suffer from performance and efficiency bottlenecks. In this research, we are incorporating blockchain technology into the FPGA and ASIC microelectronic supply chain to help mitigate the risk of counterfeit microelectronics through a secure and decentralized solution …

Design And Development Of An Immersive Simulation For Social Determinants Of Health Training, Lahari Surapaneni

Browse all Theses and Dissertations

This thesis research project focuses on design and development of an immersion simulation-based training tool that help raise the social determinants of health (SDOH) awareness among the health care providers. Compared to existing classroom lecture and/or role-play based SDOH education approach, our immersion-simulation based approach provides an easy access and highly realistic experience to such training curriculum at anytime and anywhere with an Internet connection. Such an interactive and immersive exposure is critical to raise SDOH awareness and maintain long-lasting empathy towards actual patients in practice, and thus help providers to be better prepared when encountering with those patients. Particularly, …

Data-Driven And Knowledge-Based Strategies For Realizing Crowd Wisdom On Social Media, Shreyansh Bhatt

Browse all Theses and Dissertations

The wisdom of the crowd is a well-known example of collective intelligence wherein an aggregated judgment of a group of individuals is superior to that of an individual. The aggregated judgment is surprisingly accurate for predicting the outcome of a range of tasks from geopolitical forecasting to the stock price prediction. Recent research has shown that participants' previous performance data contributes to the identification of a subset of participants that can collectively predict an accurate outcome. In the absence of such performance data, researchers have explored the role of human-perceived diversity, i.e., whether a human considers a crowd as a …

A Practitioner Survey Exploring The Value Of Forensic Tools, Ai, Filtering, & Safer Presentation For Investigating Child Sexual Abuse Material, Laura Sanchez, Cinthya Grajeda, Ibrahim Baggili, Cory Hall

Electrical & Computer Engineering and Computer Science Faculty Publications

For those investigating cases of Child Sexual Abuse Material (CSAM), there is the potential harm of experiencing trauma after illicit content exposure over a period of time. Research has shown that those working on such cases can experience psychological distress. As a result, there has been a greater effort to create and implement technologies that reduce exposure to CSAM. However, not much work has explored gathering insight regarding the functionality, effectiveness, accuracy, and importance of digital forensic tools and data science technologies from practitioners who use them. This study focused specifically on examining the value practitioners give to the tools …