Doctor of Philosophy (PhD), Wright State University, 2005, Computer Science and Engineering PhD

In the intersection of pattern recognition, machine learning, and evolutionary computation is a new search technique by which computers might program themselves. That technique is called genetic decision-programming. A computer can gain the ability to distinguish among the things that it needs to recognize by using genetic decision-programming for pattern discovery and concept learning. Those patterns and concepts can be easily encoded in the spines of a decision program (tree or diagram). A spine consists of two parts: (1) the test-outcome pairs along a path from the program's root to any of its leaves and (2) the conclusion in that leaf. The test-outcome pairs specify a pattern and the conclusion identifies the corresponding concept. Genetic decision-programming combines and extends discrete decision theory with the principles of genetics and natural selection. The resulting algorithm searches for those decision programs that best satisfy some user-defined criteria. Each program mates problem decompositions with subproblem solutions, and consists of overlapping spines. Those spines are manipulated by three context-sensitive operators. The context defines a subproblem and is determined by the operator's point of application within a decision program. Macro-mutation generatyes a new solution for that context; mini-mutation restructires the existing solution for that context; and spine crossover inserts another program's solution for that context. Those solutions are encoded in the spines. Thus the operators recompose, restructure and recombine spines as the search technique evolves a population of decision programs to satisfy the user-defined criteria. Genetic decision-programming overcomes the difficulties encountered when evolving decision programs with genetic programming techniques that rely on subtree crossover. Those impractical techniques require too much memory and computation. Subtree crossover exchanges random subtrees of broken spines without regard for c (open full item for complete abstract)

Committee: Mateen Rizki (Advisor) Subjects: Computer Science

MS, University of Cincinnati, 2024, Engineering and Applied Science: Aerospace Engineering

The integration of AI in autonomous vehicles has been rapidly expanding and has the potential to raise concerns about non-compliant or malicious actors. Predicting movements or strategies of these actors could provide a substantial advantage in the mitigation of such threats. In a simulated asteroids style game, capture of these actors closely resembles pursuit evasion problems in differential games. In this work, multiple evader control methods are mapped by an adaptable fuzzy modified potential field avoidance method trained via genetic algorithm. Evader routes are integrated and optimal interception points are determined by numerical methods or a fuzzy logic approach. Time delayed mines are then placed at the interception point to eliminate the evader. The fuzzy modified potential field has also been separately trained to produce highly effective avoidance within congested asteroid environments.

Committee: Kelly Cohen Ph.D. (Committee Chair); Anoop Sathyan Ph.D. (Committee Member); Donghoon Kim Ph.D. (Committee Member); Manish Kumar Ph.D. (Committee Member) Subjects: Aerospace Engineering

Doctor of Philosophy, The Ohio State University, 2024, Biomedical Sciences

This dissertation addresses the critical need for scalable variant interpretation in the diagnosis of rare genetic diseases (RGDs) by developing and validating novel computational methods for the interpretation of genome sequencing (GS) data. We introduce Clinical Assessment of Variant Likelihood Ratios (CAVaLRi), a robust algorithm that uses a modified likelihood ratio framework to prioritize diagnostic germline variants. CAVaLRi effectively integrates phenotypic data with variant impact predictions and family genotype information to achieve superior performance over existing prioritization tools in multiple clinical cohorts. CAVaLRi-informed reanalysis was able to uncover eight diagnoses in a cohort of RGD patients who had previously received non-diagnostic GS test results, demonstrating utility in ending diagnostic odysseys. Complementing CAVaLRi, we developed CNVoyant, an advanced tool designed to classify and prioritize copy number variants (CNVs) by incorporating machine learning techniques with genomic features to identify disease causal CNVs. CNVoyant's integration into the CAVaLRi framework allows for a unified approach to handle multiple variant types, thus providing a comprehensive solution for genetic diagnostics. The combined utility of CAVaLRi and CNVoyant offers significant improvements in diagnostic yield and accuracy, facilitating timely and precise genetic diagnosis in clinical settings. These tools represent a scalable approach to meet the growing demands of GS testing, thereby expediting the diagnostic process for patients with undiagnosed RGDs and supporting the broader application of genomics in personalized medicine.

Committee: Peter White (Advisor); Bimal Chaudhari (Advisor); Elaine Mardis (Committee Member); Alex Wagner (Committee Member); James Blachly (Committee Member) Subjects: Bioinformatics; Biomedical Research; Genetics

Doctor of Philosophy, University of Akron, 2024, Electrical Engineering

An axial flux permanent magnet (AFPM) motor can be considered one of the optimal motor designs due to its higher torque density and compact sizing. To date, most AFPM models have adopted surface mounted and spoke type rotor designs for various applications. However, these designs provide poor protection for the magnet. Furthermore, the low saliency ratio of the surface mounted AFPM motors limits their performance in the flux weakening region. On the other hand, magnets in an AFPM can be placed inside well-designed flux barriers to guard against external forces. The advantages of the proposed design can be described in the following ways. First, the flux barriers provide protection for magnets in harsh operating conditions. Second, this rotor structure generates a satisfactory level of reluctance torque which extends the operating region of the motor. Third, the simpler flux barrier simplifies the manufacturing process of the motor and reduces the manufacturing cost. This paper describes the design, optimization and prototype development of a double stator single rotor (DSSR) axial flux interior permanent magnet (AFIPM) motor with a novel “H” shaped flux barrier. Once the design parameters of the rotor and stator have been pointed out, an initial design of the proposed AFIPM has been developed in finite element analysis (FEA) based on the machining requirements and design specifications. However, the initial AFIPM design exhibited lower average torque and higher ripple. As a result, a novel multistage optimization method has been developed to achieve the desired electromagnetic performance. This optimization method, which includes Taguchi orthogonal array, multivariate regression analysis and a genetic algorithm, calculates the optimal design parameters of the motor. Detailed electromagnetic finite element analysis has been executed to compare the performances of the optimized model with the benchmark design. An AFIPM prototype has been developed to verify th (open full item for complete abstract)

Committee: Dr. Yilmaz Sozer (Advisor); Dr. J. Alexis De Abreu Garcia (Committee Member); Dr. Igor Tsukerman (Committee Member); Dr. Xiaosheng Gao (Committee Member); Dr. J. Patrick Wilber (Committee Member) Subjects: Electrical Engineering

Doctor of Philosophy (PhD), Ohio University, 2024, Industrial and Systems Engineering (Engineering and Technology)

With the onset of technology-driven solutions, the warehousing and logistics sectors are witnessing transformative advancements, one of which is the Robotic Compact Storage and Retrieval System (RCSRS). This research presents a comprehensive examination of RCSRS through three interrelated chapters. The first chapter provides an exhaustive literature review, presenting existing findings and gaps in different types of automated storage systems that have been studied, comparing their characteristics, similarities, and differences. The second chapter pioneers the study of robot travel time within RCSRS, introducing an innovative Mixed-Integer Non-Linear Programming (MINLP) model optimized using a Genetic Algorithm (GA) approach. This investigation primarily provides insights like the optimal placement of the Input/Output (I/O) point and the significance of digging time as a critical bottleneck, while also setting the stage for future research directions. Lastly, the third chapter studies the performance of three optimization algorithms in the RCSRS context: Genetic Algorithm (GA), Simulated Annealing (SA), and a novel Greedy Heuristic. This study aims to minimize robot bin moves, recognizing its direct impact on time and energy utilization. Remarkable findings such as the Greedy Heuristic's efficiency for moderate-sized order lists and the SA's aptness for larger order lists have been detailed. Together, these chapters offer an expansive view into RCSRS's potential and the strategies to harness it, contributing valuable insights and methodologies for the warehousing and logistics sectors. The research anticipates fostering advanced RCSRS designs, optimizing operations, and guiding future research in this transformative domain.

Committee: Tao Yuan Dr. (Committee Chair); Dale Masel Dr. (Committee Co-Chair); Vardges Melkonian Dr. (Committee Member); Ashley Metcalf Dr. (Committee Member); Aros-Vera Felipe Dr. (Committee Member) Subjects: Engineering; Robotics

MS, University of Cincinnati, 2023, Engineering and Applied Science: Aerospace Engineering

Fixed-wing Unmanned Aerial Vehicles (UAVs) cannot fly at speeds lower than critical stall speeds. As a result, hovering during a potential collision scenario, like with rotary-wing UAVs, is impossible. Moreover, hovering is not an optimal solution for Collision Avoidance (CA), as it increases mission time and is innately fuel inefficient. This work proposes a decentralized Fuzzy Inference System (FIS)-based resolution algorithm that modulates the point-to-point mission path while ensuring the continuous motion of UAVs during CA. A simplified kinematic guidance model with coordinated turn conditions is considered to control the UAVs. The model employs a proportional-derivative control of commanded airspeed, bank angle, and flight path angle. The commands are derived from the desired path, characterized by airspeed, heading, and altitude. The desired path is, in turn, obtained using look-ahead points generated for the target point. The FIS aims to mimic human behavior during collision scenarios, generating modulation parameters for the desired path to achieve CA. Notably, it is also scalable, which makes it easy to adjust the algorithm parameters, as per the required missions, and factors specific to a given UAV. A genetic algorithm was used to optimize FIS parameters so that the distance traveled during the mission was minimized despite path modulation. The proposed algorithm was optimized using a pairwise conflict scenario. The effectiveness of the algorithm was evaluated through various pairwise conflict scenarios as well as a Monte Carlo simulation of random conflict scenarios involving multiple UAVs operating in a confined space. It was found that the overall number of collisions decreased by an average of 98% using the proposed optimized algorithm, thereby, supporting its effectiveness.

Committee: Donghoon Kim Ph.D. (Committee Chair); Daegyun Choi Ph.D. (Committee Member); Anoop Sathyan Ph.D. (Committee Member); Ou Ma Ph.D. (Committee Member); Kelly Cohen Ph.D. (Committee Member) Subjects: Aerospace Engineering

MS, University of Cincinnati, 2023, Engineering and Applied Science: Mechanical Engineering

The accurate classification of data holds immense value across numerous domains. Although AI techniques used for classification are very accurate, they are often seen as "black boxes" that lack transparency and interoperability to explain the model. To address this limitation, this research explores the application of a cascading genetic-fuzzy system to enhance the explainability of the classification process. The focus of this master's thesis is the classification of two distinct types of rice and raisin varieties. Leveraging computer vision techniques, grayscale images of the rice samples are converted and processed to extract relevant attributes. These attributes serve as inputs to a fuzzy inference system (FIS) employing a 7-input 2-output architecture. To optimize the FIS's performance, a Genetic Algorithm (GA) is employed to fine-tune the membership functions for each input and output, as well as the rule base. The proposed cascading genetic-fuzzy system is rigorously evaluated and compared against existing methods to ascertain its effectiveness. By incorporating genetic algorithms within fuzzy systems, the approach strikes a balance between accuracy and transparency, allowing users to gain deeper insights into the classification process. Notably, the implemented system achieves an impressive accuracy of 95% and 87% on a validation set of rice and raisins respectively compared to the popular AI techniques such as Linear Regression, Standard Vector Machine, and Multi-Layer Perceptron. It is to note that cascading genetic-fuzzy system outperforms other models with more explainability. Through this master's thesis, we aim to advance the development of more interpretable and accurate classification models for real-world scenarios. By highlighting the importance of incorporating explainability into AI techniques, this research contributes to the overall understanding and utilization of transparent AI systems. The findings underscore the necess (open full item for complete abstract)

Committee: Manish Kumar Ph.D. (Committee Chair); Kelly Cohen Ph.D. (Committee Member); Sam Anand Ph.D. (Committee Member) Subjects: Mechanical Engineering

Doctor of Philosophy, The Ohio State University, 2022, Industrial and Systems Engineering

Thin-walled frames are prevalent in automotive body structures as they provide lower vehicle weight to meet strength and stiffness requirements for different types of loading cases. The inner and outer styling surfaces drive the shape of the design domain, which constrains structural configurations, and the size and shape of parts. With an increasing demand for a shorter development and production time and cost of automotive production and to stay competitive in the field, automotive manufacturers are using advanced computation techniques like Topology Optimization (TO) to create new and better performing designs. Although automotive structural engineers have been using thin-walled frames for years, they are keen on getting additional improvements through TO. However, current technology is not capable of producing thin-walled shapes. The research presented here aims to bridge the gap between the topology optimized automotive body structures and the traditional manufacturing processes by automatically post-processing the results to create hollow cross-sections for the body components to reside within a given space. The inner and outer styling surfaces control the space boundaries that constrain the structural configurations and the size of parts. Topology optimization is performed on an FEA model that represents the design domain using voxel elements, and the results typically mesh/triangulated surfaces stored as *.stl files. The objective is to develop a systematic approach for converting organic, solid, and monolithic shapes, obtained from TO, into parameterized CAD models for thin-walled structures produced by sheet metal stamping. The methods developed can perform concurrent shape and size optimization of the cross-section of thin-walled components, as opposed to the traditional method where the shape of the cross-section is obtained from an existing, which is then optimized for size. The area and moment of inertias of the cross-section are the two metrics use (open full item for complete abstract)

Committee: Jami Shah (Advisor); Ali Nassiri (Committee Member); Farhang Pourboghrat (Committee Member); Shawn Midlam-Mohler (Committee Member) Subjects: Mechanical Engineering

PhD, University of Cincinnati, 2021, Engineering and Applied Science: Civil Engineering

The primary energy sources in commercial buildings are electricity that accounts for 61%, followed by 32% for natural gas. According to EIA, the heating, ventilation, and air condition systems account for about 25% of the total commercial building's energy use in the US. Therefore, advanced modeling and optimization methods of the system components and operation offer great ways to reduce energy consumption. Since HVAC systems modeling is a characteristic and challenging process thus, while developing an HVAC system and component model, close attention should be given to the accuracy of the model structure, model parameters, and constraints. So, the final selected model can accurately deal with constraints, uncertainties, control the time-varying applications and handle a broad range of operating conditions. Also, the use of the optimization process to automate selecting the best model structure is crucial. Because every component is different, we cannot propose one model to fit that specified component in all systems. Choosing the best model structure is a time-consuming process. Here comes the optimization process role in automating the process of selecting the optimal model structure for each application. This research will introduce an innovative method of modeling and optimizing HVAC system operation to reduce the total energy consumption while improving the indoor thermal comfort level. The data-driven two-level optimization technique introduced in this research will utilize the use of real system performance data collected from the building automation systems (BAS) to create accurate component modeling and optimization process as the first level of optimization (MLO). Accurate component modeling techniques are crucial for the results accuracy of the process of optimization the HVAC systems performance. Lastly, artificial neural network (ANN) was chosen as the component modeling tool. The second level of optimization utilizes the whole system-level opt (open full item for complete abstract)

Committee: Nabil Nassif (Committee Chair); Hazem Elzarka Ph.D. (Committee Member); Amanda Webb (Committee Member); Munir Nazzal Ph.D. (Committee Member); Raj Manglik Ph.D. (Committee Member) Subjects: Engineering; Labor Economics; Theater

MS, University of Cincinnati, 2021, Engineering and Applied Science: Mechanical Engineering

Odometry in robotics is the concept of estimating the change in position of a robot with respect to some known position or origin. Precise odometry information can serve as a key component in any autonomous navigation, path planning, and map building applications. It is well known that the position estimate from the wheel odometry sensor is highly noisy due to wheel slippage. This has created the need for using different exteroceptive sensors like sonar, LiDAR, camera, etc. for odometry information. 2-D LiDAR is one of the popular sensors used for odometry estimation in the field of mobile robotics. LiDAR sensor measures time taken for reflection of the laser ray from surroundings to scan the environment. Scan matching algorithms can then be used for finding overlapping information between consecutive LiDAR scans for estimating the pose of the robot. This research proposes an algorithm based on this technique for estimating the pose of a mobile robot. Instead of using raw LiDAR scan, feature-based matching is used in which geometric features are extracted from the scan and matched across consecutive scans for estimating the pose of the robot. The proposed algorithm uses sequential RANdom SAmple Consensus (RANSAC) for extracting line features in the environment and a meta-heuristic Genetic Algorithm for scan matching. Extended Kalman Filter (EKF) has also been implemented for fusing the odometry information from the LiDAR sensor with Inertial Measurement Unit (IMU) sensor to further improve the odometry estimate of the robot. The proposed algorithm has been developed for Robot Operating System (ROS) for real-time processing making it convenient to be implemented on an actual hardware system. The algorithm has been tested in a simulation environment with experiments also performed on an actual LiDAR sensor.

Committee: Manish Kumar Ph.D. (Committee Chair); David Thompson (Committee Member); Rajnikant Sharma Ph.D. (Committee Member) Subjects: Robots

Doctor of Philosophy (PhD), Ohio University, 2021, Industrial and Systems Engineering (Engineering and Technology)

Online grocery shopping is one of the trends that became popular especially during pandemic when customers preferred to not be physically present in the stores. Many grocery stores provide the option for customers to order their products online and have them delivered to their locations. Managing the operations necessary to deliver the orders creates many new issues for the stores. There are two main factors to consider when evaluating how to provide the delivery service: minimizing the delivery cost to remain competitive with other stores and minimizing the waiting time for customers to receive their orders to maintain customer satisfaction. These factors have some commonality, in that efficient delivery routes can reduce both costs and customer waiting time. The first part of this dissertation presents a methodology for considering both cost and customer waiting time when planning delivery of grocery orders. The methodology considers customers' locations, orders' preparation time and the number of available vehicles in the store. First, it develops the method for calculating the network driving time between all locations. Then, it uses an integer programming model to develop the delivery routes. Different batching strategies are evaluated to determine whether it is preferable to dispatch vehicles quickly or to allow orders to accumulate and produce more efficient routes. Different heuristics are developed to apply these strategies on group of customers. In the second part of the research, a genetic algorithm is developed to apply on larger numbers of customers and develop the routes more quickly. The testing is performed on different customer groups by considering different operational conditions. The result of this testing is used to analyze the effect of factors like order arrival rate and the number of orders in each delivery on average driving time and average waiting time. The final part of this research applies a profit optimization model. The mo (open full item for complete abstract)

Committee: Dale Masel (Advisor); Tao Yuan (Committee Member); Saeed Ghanbartehrani (Committee Member); Vardges Melkonian (Committee Member); Mohammed Khurrum Bhutta (Committee Member) Subjects: Industrial Engineering; Information Technology; Management; Transportation Planning

MS, University of Cincinnati, 2021, Engineering and Applied Science: Computer Science

Electric power is one of the most important energy sources in the world. The stable supply of electric power plays an important role in production development, work, and life. People in all walks of life can do nothing without electricity. Therefore, if the power system is unstable, resulting in the occurrence of power failure, a certain area will be basically paralyzed, the communication will be blocked, the production will not be able to proceed, and the hospital will not be able to treat the patients, people's lives cannot be guaranteed. Therefore, it is very important to ensure the continuous, reliable and stable supply of electricity. In recent years, with the continuous development of global electricity market reform, electricity has become a freely traded commodity, and its price changes in real-time. Therefore, electricity price has become the most concerned issue. Generally, load and power generation are affected by meteorological factors such as temperature, wind speed, and precipitation, as well as the intensity of business and daily activities, such as weekends, hour etc. Therefore, electricity prices show seasonal and highly complex volatility in different time scales (daily, weekly, and annual), and there are often sudden and short-term price spikes. The fluctuation of electricity prices makes it more difficult to predict the behavior of participants in the power market and increases the risk of imbalance between supply and demand in the power market, which affects the stability of power grid operation. In the market-oriented environment of power trade, accurate price forecasting is of great significance to all stakeholders in the power market. As the buyer of the electricity trading market, it is more likely to obtain more profits in electricity trading by obtaining accurate price information in advance. From the perspective of power consumers, such as some factories with large power consumption, electricity price occupies a key part of their pr (open full item for complete abstract)

Committee: Yizong Cheng Ph.D. (Committee Chair); Ali Minai Ph.D. (Committee Member); Boyang Wang (Committee Member) Subjects: Computer Science

MS, University of Cincinnati, 2021, Engineering and Applied Science: Aerospace Engineering

Multi-drone slung payload systems that do not rely upon a leader for control instruction are a highly resilient option for critical delivery missions. In comparison to traditional control and modern deep-learning approaches, fuzzy systems are uniquely suited to balancing highly complex behaviors and high user readability post-training. By properly categorizing, training, and stacking purpose-driven fuzzy inference system (FIS) modules with separate cost functions or summed cost function components, unique behaviors can be developed and combined to form a powerful overall controller. In general, these have been referred to as genetic fuzzy trees (GFTs). Though technically the GFT proposed in this paper is a parallel set of four primary behaviors, the core program developed as a result of this research can and will easily add more tree-like series FIS modules to improve the responsiveness of the system to physical parameters. As a furthering of research in the field of genetic fuzzy decentralized control tasks, an early solution is explored for full three-dimensional point-mass control of a payload with neighbor avoidance safety and basic linear targeting behaviors. Drone teams of nominal counts (three to four drones depending on training task) are attached to a point mass payload by identical cables and are provided with visually obtainable state information. Isolated trainings build behaviors in tasks including cable pitch, altitude, cable yaw, and planar navigation, with a focus on visual representation of learned behaviors via fuzzy rule base surface plots. In response to difficulties experienced in setup of prior two-dimensional tests, a highly customizable object-oriented architecture dubbed GRAFT (Genetic ReArrangeable Fuzzy Tuner) for both simulation trial and fuzzy inference system creation has been developed and is explored briefly as a vital tool for future hand-built multi-inference system fuzzy controllers. This design (open full item for complete abstract)

Committee: Ou Ma Ph.D. (Committee Chair); Kelly Cohen Ph.D. (Committee Member); Catharine McGhan Ph.D. (Committee Member); Anoop Sathyan PhD (Committee Member) Subjects: Aerospace Materials

Master of Science in Electrical Engineering, University of Dayton, 2020, Electrical and Computer Engineering

The purpose of this thesis is to produce a closed-loop feedback mission planning tool that allows for the operator to control multiple Unmanned Aerial Vehicles (UAV) within a mission. Different styles of UAVs and mission planners that are available on the market were evaluated and selected for their cost, size, ability to customize, and fit for mission work. It was determined that commercially available mission planners do not provide the level of automation required, such as allowing for different algorithms for assigning UAV tasks and for planning UAV flight paths within a mission. Comparisons were made between different algorithms for path planning and tasking. From these comparisons, a bio-inspired machine-learning algorithm, Genetic Algorithm (GA), was chosen for assigning tasks to UAVs and Dubins path was chosen for modeling UAV flight paths within the mission simulation. Since market mission planners didn't allow for control of multiple UAVs, or wouldn't allow for the operator to add algorithms to increase usability and automation of the program, it was decided to create a Graphic User Interface (GUI) that would allow the operator to customize UAVs and the mission scenario. A test mission scenario was then designed, which included 9 Points of Interest (POI), 1 to 3 Targets of Interest (TOI), 3 to 5 UAVs, as well as simulation options that modeled failure of a task or a UAV crash. Operator feedback was incorporated into the simulation by allowing the operator to determine a course of action if a failure occurred, such as reprogramming the other UAVs to complete the tasks left by the crashed UAV or reassessing a failed task. Overall mission times decreased for reprogramming the UAVs versus running a separate mission to complete any tasks left by the crashed UAV. Additional code was added to the GA and Dubins path to increase speed without decreasing solution fitness.

Committee: Amy Neidhard-Doll Ph.D. (Advisor); Eric Balster Ph.D. (Committee Member); Bradley Ratliff Ph.D. (Committee Member) Subjects: Electrical Engineering

PhD, University of Cincinnati, 2020, Engineering and Applied Science: Mechanical Engineering

Broadband active noise control (ANC) presents challenges due to both computational demand and slow convergence. Subband adaptive filter (SAF) has hence been introduced to address these challenges, especially when the broadband noise is with large spectral dynamic range and the ANC system model has long impulse response. The study of subband ANC usually requires a broad understanding of the general broadband ANC, as well as the SAF architecture including the analysis filter bank for signal decomposition, the adaptive algorithm for subband filter weights adaptation, and the weight stacking for combining subband filter weights into a fullband noise cancelling filter. A more in-depth understanding of a general broadband ANC is needed to help predict the potentially maximum attenuation level of a subband ANC, and guide the way how the system could be effectively initialized for rapid convergence. The design of SAF architecture for broadband ANC remains very challenging, which needs a trade-off between various effects including delay and spectral leakage due to analysis filter bank, computational cost of subband adaptive algorithm, and weight stacking distortion. This necessitates an effective method to find out an optimal parametric design solution for the subband ANC. In practice, many unwanted interferences and high-amplitude disturbances may threaten the effectiveness, stability and robustness of subband ANC, which has not been thoroughly investigated. Thus, we develop an effective and computationally efficient subband ANC system with robust adaptive control algorithms for broadband noise. First, this dissertation presents the maximum achievable noise reduction of a general feedforward broadband ANC system corresponding to the optimal filter under causality constraint. This optimal causal filter helps guide the initialization of the filter with finite length, and guarantees rapid convergence and maximum attenuation performance. Besides theoretical analysis, numeric (open full item for complete abstract)

Committee: Jay Kim Ph.D. (Committee Chair); Manish Kumar Ph.D. (Committee Member); Teik Lim Ph.D. (Committee Member); Yongfeng Xu Ph.D. (Committee Member) Subjects: Mechanical Engineering

Doctor of Philosophy, University of Akron, 2020, Integrated Bioscience

The Gaussian graphical model (GGM) is widely applied to learn genetic network since it defines an undirected graph decoding the conditional dependence between genes. Many algorithms based on the GGM have been proposed for learning genetic network structures. Since the number of gene variables is typically far more than the number of samples collected, and a real genetic network is typically sparse, the graphical lasso implementation of GGM becomes a popular tool for inferring the conditional interdependence among genes. In this study, based on the guidance of specific types of human cancer pathways in the Kyoto Encyclopedia of Genes and Genomes (KEGG), I extracted the genes involved in a specific KEGG pathway and the corresponding RNA-seq expression levels in cancer and normal tissues from The Cancer Genome Atlas (TCGA), and constructed two types of small gene expression datasets: normal and cancer gene expression datasets corresponding to gene sets of different types of human cancers. I directly applied graphical lasso to the gene expression datasets of the genes to infer their genetic conditional dependences. By integrated analysis and comparison on these inferred normal and cancer networks, the results reveal highly conditional dependences among the genes at the RNA-seq expression levels and further confirm the essential roles played by the genes that encode proteins involved in the two-key signaling pathways phosphoinositide 3-kinase (PI3K)/AKT/mTOR and Ras/Raf/MEK/ERK in human carcinogenesis. These highly conditional dependences elucidate the expression level interactions among the genes that are implicated in many different human cancers. The inferred genetic networks were examined to further identify and characterize a collection of gene interactions that are unique to cancers. The cross-cancer genetic interactions revealed from our study provide another set of knowledge for cancer biologists to propose strong hypotheses, so further biological investi (open full item for complete abstract)

Committee: Zhong-Hui Duan (Advisor); Sujay Datta (Committee Member); Qin Liu (Committee Member); Timothy O'Neil (Committee Member); Yingcai Xiao (Committee Member) Subjects: Bioinformatics; Computer Science

PhD, University of Cincinnati, 2020, Engineering and Applied Science: Aerospace Engineering

A novel approach to parameterize and solve for optimal satellite attitude state trajectories is presented. The optimal trajectories are parameterized using fuzzy inference systems (FISs), and the FISs are optimized using a genetic algorithm. Eight different constrained optimization problems are solved. The objective of each optimization problem is either battery charge maximization, link margin (equivalent to antenna gain) maximization, or experiment temperature minimization. All optimization problems consider reaction wheel angular velocity and reaction wheel angular acceleration constraints, and five of the optimization problems consider either battery charge constraints, antenna gain constraints, or both battery charge and antenna gain constraints. Reaction wheel constraints are satisfied using an attitude state filter at the output of the FISs and an optimal magnetic torque / reaction wheel desaturation algorithm, the design of both of which is presented herein. Optimal attitude state trajectory, or attitude profile, FISs are compared with a nominal attitude profile. It is shown that, while the nominal attitude profile offers good performance with respect to both battery charge and link margin, the optimal attitude profile FISs are able to outperform the nominal profile with respect to all objectives, and a minimum temperature attitude profile FIS is able to achieve average experiment temperatures 30–40 K lower than the nominal attitude profile. The attitude state trajectory optimization solutions presented in this work are motivated by the needs and constraints of the CryoCube-1 mission. Because this work is integral to the functionality of the CryoCube-1 satellite system, the effort taken to successfully build, test, deliver, launch, and deploy this CubeSat is detailed. The intent of providing this systems view is to provide the context necessary to understand exactly how the attitude state trajectory optimization results were used within the satellite system.

Committee: Kelly Cohen Ph.D. (Committee Chair); Manish Kumar Ph.D. (Committee Member); Ou Ma Ph.D. (Committee Member); Phil Putman Ph.D. (Committee Member); Anoop Sathyan Ph.D. (Committee Member) Subjects: Aerospace Materials

Master of Science in Computer Engineering (MSCE), Wright State University, 2019, Computer Engineering

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 search space reduction methodology is effective in producing an error correcting oscillator configuration rapidly, online, while the vehicle is in normal service.

Committee: John C. Gallagher Ph.D. (Advisor); Michael L. Raymer Ph.D. (Committee Member); Mateen Rizki Ph.D. (Committee Member) Subjects: Computer Engineering

MS, University of Cincinnati, 2019, Engineering and Applied Science: Computer Science

As the applications of Unmanned Aerial Vehicles (UAVs) are becoming more and more common, it is necessary to address their inherent technological challenges so as to make them safe and more useful. Designing an eective UAV path planning algorithm is essential in all UAV missions. The UAV path planning strategy depends on its application eld. The application specic constraints also need to be satised along with UAV mobility aspects for a successful UAV mission. This thesis aims to solve the UAV Path planning problem for the scenarios when the UAVs are used for remote sensing and data communication applications. The thesis consists of two pieces of work. The rst piece of work addresses and solves a UAV path planning problem when time windows and data ooading constraints are involved, which is very typical in a data communication application scenario. A Genetic Algorithm based approach is used to solve the problem in realistic time limits. In the second piece of work, an evaluation of the capabilities of currently available robotics and network simulators is conducted to determine their suitability to be used as a simulator for multimedia data communication over UAV networks. An ideal simulator for this purpose should have simulation capabilities for image/video capture, image processing, encoding/decoding and quality measurement along with flight and network simulation. A new simulation framework is proposed and tested by combining X-Plane, M3WSN and EvalVid simulator platforms to achieve an end to end simulation of a UAV multimedia data communication scenario.

Committee: Rui Dai Ph.D. (Committee Chair); Dharma Agrawal D.Sc. (Committee Member); Manish Kumar Ph.D. (Committee Member) Subjects: Computer Science

Doctor of Philosophy (Ph.D.), University of Dayton, 2018, Aerospace Engineering

So-called wind-lens turbines offer the potential for improved energy efficiency and better suitability for urban and suburban environments compared to unshrouded or bare wind turbines. Wind-lenses, which are typically comprised of a diffuser shroud equipped with a flange, can enhance the wind velocity at the rotor plane due to the generation of a lower back pressure. This work comprises of two main studies which aim to develop fast and accurate simulation tools for the performance prediction and design of shrouded horizontal axis wind turbines. In the first study, a low-order theoretical model of ducted turbines is developed to establish a better understanding of the basic aerodynamics of shrouded wind turbines. Then a cost-effective CFD tool coupled with a multi-objective genetic algorithm is developed and employed to improve the performance of shrouded wind turbines. A low-order semi empirical model, which offers performance prediction for the power and thrust coefficients, is developed and applied to shrouded turbines. This 1D model is based on assumptions and approximations to calculate optimal power coefficients and power extraction, as well as augmentation ratios. It is revealed that the power enhancement is proportional to the mass stream rise produced by the nozzle diffuser-augmented wind turbine (NDAWT). Such mass flow rise can only be accomplished through two essential principles: an increase in the area ratios and/or by reducing the negative back pressure at the exit. The thrust coefficient for optimal power production of a conventional bare wind turbine is known to be 8/9, whereas the theoretical analysis of the NDAWT predicts an ideal thrust coefficient either lower or higher than 8/9 depending on the back-pressure coefficient at which the shrouded turbine operates. Computed performance expectations demonstrate a good agreement with numerical and experimental results, and it is demonstrated that much larger power coefficients than for traditional win (open full item for complete abstract)

Committee: Markus Rumpfkeil Ph.D. (Advisor); Kevin Hallinan Ph.D. (Committee Member); Andrew Chiasson Ph.D. (Committee Member); Youssef Raffoul Ph.D. (Committee Member) Subjects: Aerospace Engineering; Mechanical Engineering