Theses and Dissertations from UMD
Permanent URI for this communityhttp://hdl.handle.net/1903/2
New submissions to the thesis/dissertation collections are added automatically as they are received from the Graduate School. Currently, the Graduate School deposits all theses and dissertations from a given semester after the official graduation date. This means that there may be up to a 4 month delay in the appearance of a give thesis/dissertation in DRUM
More information is available at Theses and Dissertations at University of Maryland Libraries.
Browse
4 results
Search Results
Item Deployment of Large Vision and Language Models for Real-Time Robotic Triage in a Mass Casualty Incident(2024) Mangel, Alexandra Paige; Paley, Derek; Aerospace Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)In the event of a mass casualty incident, such as a natural disaster or war zone, having a system of triage in place that is efficient and accurate is critical for life-saving intervention, but medical personnel and resources are often strained and struggle to provide immediate care to those in need. This thesis proposes a system of autonomous air and ground vehicles equipped with stand-off sensing equipment designed to detect and localize casualties and assess them for critical injury patterns. The goal is to assist emergency medical technicians in identifying those in need of primary care by using generative AI models to analyze casualty images and communicate with the victims. Large language models are explored for the purpose of developing a chatbot that can ask a casualty where they are experiencing pain and make an informed assessment about injury classifications, and a vision language model is prompt engineered to assess a casualty image to produce a report on designated injury classifiers.Item A Binary Classifier for Test Case Feasibility Applied to Automatically Generated Tests of Event-Driven Software(2016) Robbins, Bryan Thomas; Memon, Atif; Computer Science; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)Modern software application testing, such as the testing of software driven by graphical user interfaces (GUIs) or leveraging event-driven architectures in general, requires paying careful attention to context. Model-based testing (MBT) approaches first acquire a model of an application, then use the model to construct test cases covering relevant contexts. A major shortcoming of state-of-the-art automated model-based testing is that many test cases proposed by the model are not actually executable. These \textit{infeasible} test cases threaten the integrity of the entire model-based suite, and any coverage of contexts the suite aims to provide. In this research, I develop and evaluate a novel approach for classifying the feasibility of test cases. I identify a set of pertinent features for the classifier, and develop novel methods for extracting these features from the outputs of MBT tools. I use a supervised logistic regression approach to obtain a model of test case feasibility from a randomly selected training suite of test cases. I evaluate this approach with a set of experiments. The outcomes of this investigation are as follows: I confirm that infeasibility is prevalent in MBT, even for test suites designed to cover a relatively small number of unique contexts. I confirm that the frequency of infeasibility varies widely across applications. I develop and train a binary classifier for feasibility with average overall error, false positive, and false negative rates under 5\%. I find that unique event IDs are key features of the feasibility classifier, while model-specific event types are not. I construct three types of features from the event IDs associated with test cases, and evaluate the relative effectiveness of each within the classifier. To support this study, I also develop a number of tools and infrastructure components for scalable execution of automated jobs, which use state-of-the-art container and continuous integration technologies to enable parallel test execution and the persistence of all experimental artifacts.Item PLANNING FOR AUTOMATED OPTICAL MICROMANIPULATION OF BIOLOGICAL CELLS(2013) CHOWDHURY, SAGAR; Gupta, Satyandra K.; Mechanical Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)Optical tweezers (OT) can be viewed as a robot that uses a highly focused laser beam for precise manipulation of biological objects and dielectric beads at micro-scale. Using holographic optical tweezers (HOT) multiple optical traps can be created to allow several operations in parallel. Moreover, due to the non-contact nature of manipulation OT can be potentially integrated with other manipulation techniques (e.g. microfluidics, acoustics, magnetics etc.) to ensure its high throughput. However, biological manipulation using OT suffers from two serious drawbacks: (1) slow manipulation due to manual operation and (2) severe effects on cell viability due to direct exposure of laser. This dissertation explores the problem of autonomous OT based cell manipulation in the light of addressing the two aforementioned limitations. Microfluidic devices are well suited for the study of biological objects because of their high throughput. Integrating microfluidics with OT provides precise position control as well as high throughput. An automated, physics-aware, planning approach is developed for fast transport of cells in OT assisted microfluidic chambers. The heuristic based planner employs a specific cost function for searching over a novel state-action space representation. The effectiveness of the planning algorithm is demonstrated using both simulation and physical experiments in microfluidic-optical tweezers hybrid manipulation setup. An indirect manipulation approach is developed for preventing cells from high intensity laser. Optically trapped inert microspheres are used for manipulating cells indirectly either by gripping or pushing. A novel planning and control approach is devised to automate the indirect manipulation of cells. The planning algorithm takes the motion constraints of the gripper or pushing formation into account to minimize the manipulation time. Two different types of cells (Saccharomyces cerevisiae and Dictyostelium discoideum) are manipulated to demonstrate the effectiveness of the indirect manipulation approach.Item Real-Time Path Planning for Automating Optical Tweezers based Particle Transport Operations(2009) Banerjee, Ashis Gopal; Gupta, Satyandra K; Mechanical Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)Optical tweezers (OT) have been developed to successfully trap, orient, and transport micro and nano scale components of many different sizes and shapes in a fluid medium. They can be viewed as robots made out of light. Components can be simply released from optical traps by switching off laser beams. By utilizing the principle of time sharing or holograms, multiple optical traps can perform several operations in parallel. These characteristics make optical tweezers a very promising technology for creating directed micro and nano scale assemblies. In the infra-red regime, they are useful in a large number of biological applications as well. This dissertation explores the problem of real-time path planning for autonomous OT based transport operations. Such operations pose interesting challenges as the environment is uncertain and dynamic due to the random Brownian motion of the particles and noise in the imaging based measurements. Silica microspheres having diameters between (1-20) µm are selected as model components. Offline simulations are performed to gather trapping probability data that serves as a measure of trap strength and reliability as a function of relative position of the particle under consideration with respect to the trap focus, and trap velocity. Simplified models are generated using Gaussian Radial Basis Functions to represent the data in a compact form. These metamodels can be queried at run-time to obtain estimated probability values accurately and efficiently. Simple trapping probability models are then utilized in a stochastic dynamic programming framework to compute optimum trap locations and velocities that minimizes the total, expected transport time by incorporating collision avoidance and recovery steps. A discrete version of an approximate partially observable Markov decision process algorithm, called the QMDP_NLTDV algorithm, is developed. Real-time performance is ensured by pruning the search space and enhancing convergence rates by introducing a non-linear value function. The algorithm is validated both using a simulator as well as a physical holographic tweezer set-up. Successful runs show that the automated planner is flexible, works well in reasonably crowded scenes, and is capable of transporting a specific particle to a given goal location by avoiding collisions either by circumventing or by trapping other freely diffusing particles. This technique for transporting individual particles is utilized within a decoupled and prioritized approach to move multiple particles simultaneously. An iterative version of a bipartite graph matching algorithm is also used to assign goal locations to target objects optimally. As in the case of single particle transport, simulation and some physical experiments are performed to validate the multi-particle planning approach.