Other Electrical and Computer Engineering Commons^™

Hardware-In-The-Loop Evaluation Of Sensor-Source Selection For Prosthetic Locomotion Intent Recognition, Victoria Asencio-Clemens

Master's Theses

Active lower-limb prostheses use intent-recognition systems to identify a user’s locomotion mode and select an appropriate control strategy, but sensor configurations that perform well offline may be unsuitable for resource-constrained embedded hardware. Existing sensor-selection methods generally prioritize classification accuracy without directly accounting for processing latency, memory usage, or other hardware-dependent requirements. To address this limitation, this thesis develops a hardware-in-the-loop source-selection framework for embedded classification of level walking, ramp ascent, ramp descent, stair ascent, and stair descent using multimodal biomechanical data from transtibial amputee participants. Subject-specific linear support vector machine classifiers were evaluated using trial-held-out validation, and candidate configurations from …

Design And Parametric Study Of A Mems-Based Reservoir Computer For Reinforcement Learning, Andrew P. Carr

Master's Theses

Single-node reservoir computing (RC) is a hardware-efficient approach to machine learning, leveraging the dynamics of physical systems. In this work, two reinforcement learning algorithms, Q-learning and Proximal Policy Optimization (PPO), are applied to a simulated micro-electro-mechanical system (MEMS)-based reservoir computer to solve both discrete and continuous control tasks. MEMS-based reservoirs are low-power, compact, and their natural frequencies (kHz to MHz) pair well with real-time control loops. To explore the relationship between reservoir dynamics and learning performance, a parametric study is conducted on two reservoir hyperparameters, reservoir size and neuron separation, using CartPole-v1 and MountainCar-v0. The RC successfully learns multiple tasks …

Computer Vision Methods For Detecting Counterfeit Usd Bills, Tyler W. Jones

Master's Theses

This thesis addresses the global challenge of counterfeit paper currency by proposing a classical computer vision framework for distinguishing genuine United States Dollar (USD) bills from counterfeit ones using image data. In contrast to existing approaches that rely on a large number of easily reproducible visual features, this work prioritizes the detection of a single, robust security feature: the ultraviolet (UV) reactive security strip embedded in genuine USD bills of denominations $5 and above. By focusing on a feature that is inherently difficult to replicate, the proposed method reduces the likelihood of counterfeit bills being misclassified as genuine.

The system …

Design And Reliability Analysis Of A Radiation-Tolerant On-Board Computer System For Martian Surface Missions, Jack Ryan

Master's Theses

Space environments present complex challenges for electronic devices, perhaps most notably in the form of radiation effects; the natural protections provided by Earth’s atmosphere and magnetosphere are largely absent in deep space and extraterrestrial environments, making single-event effects (SEE) a critical concern. Although radiation-hardened components offer near-immunity to SEE, they possess tremendous drawbacks in both cost and performance. To circumvent such issues, this thesis investigates the feasibility of leveraging a commercial-off-the-shelf (COTS) device, the AMD KRIA K24 system-on-module (SOM), for use in Martian surface missions.

Detailed models were used to predict SEE rates in the system, and system-level fault tree …

Integration And Testing Of A Quadruped Robot With Ros2, Jeremy S. West

Master's Theses

The Cal Poly Legged Robotics Group has been developing research and teaching platforms for agile legged robotics since 2020. These platforms are expected to provide students with opportunities to develop complete legged-robot systems from low-level control to advanced robotics tasks such as motion planning and decision making. However, the current prototyped quadruped robot lacked the software and sensing capabilities for high-level quadrupedal gaits and advanced robotic research.

To address these challenges, this project developed Switch, a robotic platform that builds upon the previous BRUCE platform with significant hardware and software upgrades. Switch features a modular design that allows individual software …

Seconds From Impact: Anticipatory Vehicular Crash Prediction Using Video Vision Transormers, Ryan P. Geisen

Master's Theses

Vehicular collisions represent a significant public health concern, necessitating re search into advanced emergency notification systems. While deep learning has shown promise in accident detection, a research gap persists in applying state-of-the-art transformer architectures to the task of anticipatory, real-time crash prediction from video. This thesis addresses this gap by developing and evaluating a Video Vision Transformer (ViViT) for the binary classification of imminent vehicular collisions. Utilizing a curated dataset of 1,493 unique collision sequences, this study systemati cally investigates the impact of temporal context by comparing the ViViT against a single-frame Vision Transformer (ViT) baseline and conducting comprehensive exper …

Enhanced Post-Capture Automatic White Balance For Srgb Images, Eric Huang

Master's Theses

Automatic white balancing (AWB) aims to correct color casts caused by varying illumination conditions, typically assuming access to RAW sensor data. However, many real-world applications involve only sRGB images that have already been processed by in-camera pipelines. In these cases, traditional AWB algorithms often underperform due to the nonlinear transformations done by these pipelines.

This thesis builds upon a data-driven color correction framework introduced by Afifi et al. that relies on RGB-UV histograms and learned color transforms. A revised automatic white balancing (AWB) framework that improves both color accuracy and runtime efficiency is proposed. A fallback routine is implemented to …

Yolot: A Recurrent Yolo Model For Robust Video-Based Automotive Object Detection, Dylan Jay Baxter

Master's Theses

Though incredibly effective at detecting objects in isolated frames, modern object detection models are often not designed to take advantage of information present in previous frames of a video stream, despite that data being readily avail- able. To address this shortcoming, this paper proposes YOLOT, a modification of the widely used YOLOv8 object detection model, which seeks to utilize this temporal information with the addition of recurrent structures. In the design of YOLOT, a series of recurrent convolutional modules were inserted at backbone and neck outputs and the final and most effective design was found to be the insertion of …

Perceptual Hash Based Content Matching, Nicholas Daniel Chenevey

Master's Theses

The proliferation of video content and AI generated imagery has introduced a number of new challenges in content identification and verification. The ability to trace content back to its source has become a critical problem as video content increases both naturally and synthetically through AI generation. This thesis provides the design, analysis, and experimental verification for a perceptual hash based framework aimed at addressing these challenges. Perceptual hashing is a method for encoding the visual content of images into compact and easily comparable binary strings. This process is used as the foundation for content matching in videos and source verification …

Hypoxic Incubator: Improving Robustness/Reliability And Demonstrating Physiological Efficacy, Damon Dennis Tan

Master's Theses

The Microphysiological Systems Laboratory aims to develop colorectal cancer tumor models under a hypoxic environment to assess model response to pharmaceutical compounds in vitro. To perform relevant studies, researchers have attempted to use different hypoxic inducing strategies such as a nitrogen pod and hypoxic incubator to recreate in vivo physiological responses to hypoxia. However, studies would be interrupted due to incubator functionality failure. To ensure successful and physiologically relevant studies, I improved and verified the robustness and reliability of a hypoxic incubator previously designed and manufactured in the lab. Through the testing and iterating design processes, I engineered and implemented …

3d Printed Microfluidic Fabrication Methodology, Characterization, Mechanical Design, And Applications In Electrostatic Artificial Muscles And Benthic Microbial Fuel Cells, Terak B. Hornik

Master's Theses

The fabrication of microfluidic devices often requires specialized methods. The development of these methods requires careful characterization and understanding of the processes involved. Using primarily PolyJet 3D printing technology, microfluidics offers a wide scope of applications such as microfluidic benthic microbial fuel cells (MBMFCs) and electrostatic artificial muscles. MBMFCs benefit from the confinement of the microbes resulting in close proximity between the electrode and the organisms. Using a modular design called the Sponge, assembly and upscaling is possible. Electrostatic artificial muscles benefit from a microfluidic approach due to the non-linearity of electrostatic attraction creating disproportionate benefits when miniaturized. When designed …

A Study Of Random Partitions Vs. Patient-Based Partitions In Breast Cancer Tumor Detection Using Convolutional Neural Networks, Joshua N. Ramos

Master's Theses

Breast cancer is one of the deadliest cancers for women. In the US, 1 in 8 women will be diagnosed with breast cancer within their lifetimes. Detection and diagnosis play an important role in saving lives. To this end, many classifiers with varying structures have been designed to classify breast cancer histopathological images. However, randomly partitioning data, like many previous works have done, can lead to artificially inflated accuracies and classifiers that do not generalize. Data leakage occurs when researchers assume that every image in a dataset is independent of each other, which is often not the case for medical …

A Study On Rapidly Exploring Random Tree Algorithms For Robot Path Planning, Sahil Sharma

Master's Theses

Robot path planning is a critical feature of autonomous systems. Rapidly-exploring Random Trees (RRT) is a path planning technique that randomly samples the robot configuration space to find a path between the start and end point. This thesis studies and compares the performance of four important RRT algorithms, namely, the original RRT, the optimal RRT (also termed RRT*), RRT*-Smart, and Informed RRT* for six different environments. The performance measures include the final path length (which is also the shortest path length found by each algorithm), time to find the first path, run time (of 1000 iterations) for each algorithm, total …

Enhancing Telecom Churn Prediction: Adaboost With Oversampling And Recursive Feature Elimination Approach, Long Dinh Tran

Master's Theses

Churn prediction is a critical task for businesses to retain their valuable customers. This paper presents a comprehensive study of churn prediction in the telecom sector using 15 approaches, including popular algorithms such as Logistic Regression, Support Vector Machine, Decision Tree, Random Forest, and AdaBoost.

The study is segmented into three sets of experiments, each focusing on a different approach to building the churn prediction model. The model is constructed using the original training set in the first set of experiments. The second set involves oversampling the training set to address the issue of imbalanced data. Lastly, the third set …

A Nano-Drone Safety Architecture, Connor J. Sexton

Master's Theses

As small-form factor drones grow more intelligent, they increasingly require more sophisticated capabilities to record sensor data and system state, ensuring safe and improved operation. Already regulations for black boxes, electronic data recorders (EDRs), for determining liabilities and improving the safety of large-form factor autonomous vehicles are becoming established. Conventional techniques use hardened memory storage units that conserve all sensor (visual) and system operational state; and N-way redundant models for detecting uncertainty in system operation. For small-form factor drones, which are highly limited by weight, power, and computational resources, these techniques become increasingly prohibitive. In this paper, we propose a …

Bearing Fault Detection And Classification Using Artificial Neural Networks, Harnak Singh

Master's Theses

Bearings are the essential components of modern rotating machines. Bearing faults can cause severe machine damages or even breakdowns.

In recent years, artificial intelligence and deep learning have been successfully applied to fault detection. In this thesis, convolutional neural networks (CNN) are employed for bearing fault detection and classification. Computer simulations results demonstrate that the CNN based approach is advantageous over the conventional regression model, with an overall accuracy of 99.5%.

Investigation Of Green Strawberry Detection Using R-Cnn With Various Architectures, Daniel W. Rivers

Master's Theses

Traditional image processing solutions have been applied in the past to detect and count strawberries. These methods typically involve feature extraction followed by object detection using one or more features. Some object detection problems can be ambiguous as to what features are relevant and the solutions to many problems are only fully realized when the modern approach has been applied and tested, such as deep learning.

In this work, we investigate the use of R-CNN for green strawberry detection. The object detection involves finding regions of interest (ROIs) in field images using the selective segmentation algorithm and inputting these regions …

An Analysis Of Camera Configurations And Depth Estimation Algorithms For Triple-Camera Computer Vision Systems, Jared Peter-Contesse

Master's Theses

The ability to accurately map and localize relevant objects surrounding a vehicle is an important task for autonomous vehicle systems. Currently, many of the environmental mapping approaches rely on the expensive LiDAR sensor. Researchers have been attempting to transition to cheaper sensors like the camera, but so far, the mapping accuracy of single-camera and dual-camera systems has not matched the accuracy of LiDAR systems. This thesis examines depth estimation algorithms and camera configurations of a triple-camera system to determine if sensor data from an additional perspective will improve the accuracy of camera-based systems. Using a synthetic dataset, the performance of …

First Order Self-Oscillating Class-D Circuit With Triangular Wave Injection, Matthew J. Carroll

Master's Theses

An investigation into performance improvements to the modulator stage of a class-D amplifier is conducted in this thesis. Two of the standard topologies, namely class-D open-loop pulse-width modulation (PWM), and the improved self-oscillating feedback system are benchmarked against a topology which includes both a hysteretic comparator in a feedback loop and triangle wave injection. Circuit performance is analyzed by comparing how the triangle injection circuit handles known issues with open-loop and self-oscillating circuits. Using this analysis, it is shown that the triangle injection topology offers an improved power supply rejection ratio relative to open-loop PWM and reduces distortion generated by …

An Artificial Neural Network For Bankruptcy Prediction, Walter D. Magdefrau

Master's Theses

Assessing the financial health of organizations remains a topic of great interest to economists, financial institutions, and invested stakeholders. For more than a century, research into financial distress has focused primarily on traditional applications of statistical analysis; however, modern advances in computational efficiency have created a significant opportunity for more sophisticated approaches. This thesis investigates the application of artificial intelligence on company bankruptcy prediction. The proposed neural network model is evaluated using the Polish Companies Bankruptcy dataset and yields a 5-year prediction accuracy of 96.5% and an AUC (area under receiver operating characteristic curve) measure of 92.4%.

Biological Semantic Segmentation On Ct Medical Images For Kidney Tumor Detection Using Nnu-Net Framework, Andres Bergsneider

Master's Theses

Healthcare systems are constantly challenged with bottlenecks due to human-reliant operations, such as analyzing medical images. High precision and repeatability is necessary when performing a diagnostics on patients with tumors. Throughout the years an increasing number of advancements have been made using various machine learning algorithms for the detection of tumors helping to fast track diagnosis and treatment decisions. “Black Box” systems such as the complex deep learning networks discussed in this paper rely heavily on hyperparameter optimization in order to obtain the most ideal performance. This requires a significant time investment in the tuning of such networks to acquire …

Boost Converter Inductor Sizing Effects On The Performance Of Mppt Algorithms, Alan Nonaka

Master's Theses

With solar power and other renewables set to take over the market in the coming decades, maximum power point tracking will be essential to optimizing power output. One underserved topic of research is the effect of inductor current ripple on performance of Maximum Power Point Tracking (MPPT) algorithms. Many new topologies are focused on decreasing the ripple from PV source to increase efficiency and power output. However, not much has been done to show ripple degrading performance of MPPT algorithms. This study uses a boost converter topology to test the performance of constant duty cycle step Perturb and Observe (PO), …

Electricity Price Forecasting Using A Convolutional Neural Network, Elliott Winicki

Master's Theses

Many methods have been used to forecast real-time electricity prices in various regions around the world. The problem is difficult because of market volatility affected by a wide range of exogenous variables from weather to natural gas prices, and accurate price forecasting could help both suppliers and consumers plan effective business strategies. Statistical analysis with autoregressive moving average methods and computational intelligence approaches using artificial neural networks dominate the landscape. With the rise in popularity of convolutional neural networks to handle problems with large numbers of inputs, and convolutional neural networks conspicuously lacking from current literature in this field, convolutional …

Performance Enhancement And Characterization Of An Electromagnetic Railgun, Paul M. Gilles

Master's Theses

Collision with orbital debris poses a serious threat to spacecraft and astronauts. Hypervelocity impacts resulting from collisions mean that objects with a mass less than 1g can cause mission-ending damage to spacecraft. A means of shielding spacecraft against collisions is necessary. A means of testing candidate shielding methods for their efficacy in mitigating hypervelocity impacts is therefore also necessary. Cal Poly’s Electromagnetic Railgun was designed with the goal of creating a laboratory system capable of simulating hypervelocity (≥ 3 km/s) impacts. Due to several factors, the system was not previously capable of high-velocity (≥ 1 km/s) tests. A deficient projectile …

Utilizing Trajectory Optimization In The Training Of Neural Network Controllers, Nicholas Kimball

Master's Theses

Applying reinforcement learning to control systems enables the use of machine learning to develop elegant and efficient control laws. Coupled with the representational power of neural networks, reinforcement learning algorithms can learn complex policies that can be difficult to emulate using traditional control system design approaches. In this thesis, three different model-free reinforcement learning algorithms, including Monte Carlo Control, REINFORCE with baseline, and Guided Policy Search are compared in simulated, continuous action-space environments. The results show that the Guided Policy Search algorithm is able to learn a desired control policy much faster than the other algorithms. In the inverted pendulum …

Development Of A Myoelectric Detection Circuit Platform For Computer Interface Applications, Nickolas Andrew Butler

Master's Theses

Personal computers and portable electronics continue to rapidly advance and integrate into our lives as tools that facilitate efficient communication and interaction with the outside world. Now with a multitude of different devices available, personal computers are accessible to a wider audience than ever before. To continue to expand and reach new users, novel user interface technologies have been developed, such as touch input and gyroscopic motion, in which enhanced control fidelity can be achieved. For users with limited-to-no use of their hands, or for those who seek additional means to intuitively use and command a computer, novel sensory systems …

Logging, Visualization, And Analysis Of Network And Power Data Of Iot Devices, Neal Huynh Nguyen

Master's Theses

There are approximately 23.14 billion IoT(Internet of Things) devices currently in use worldwide. This number is projected to grow to over 75 billion by 2025. Despite their ubiquity little is known about the security and privacy implications of IoT devices. Several large-scale attacks against IoT devices have already been recorded.

To help address this knowledge gap, we have collected a year’s worth of network traffic and power data from 16 common IoT devices. From this data, we show that we can identify different smart speakers, like the Echo Dot, from analyzing one minute of power data on a shared power …

Iris Biometric Identification Using Artificial Neural Networks, Kevin Joseph Haskett

Master's Theses

A biometric method is a more secure way of personal identification than passwords. This thesis examines the iris as a personal identifier with the use of neural networks as the classifier. A comparison of different feature extraction methods that include the Fourier transform, discrete cosine transform, the eigen analysis method, and the wavelet transform, is performed. The robustness of each method, with respect to distortion and noise, is also studied.

Towards A Strawberry Harvest Prediction System Using Computer Vision And Pattern Recognition, Andreas M. Apitz

Master's Theses

Farmers require advance notice when a harvest is approaching, so they can allocate resources and hire workers as efficiently as possible. Existing methods are subjective and labor intensive, and require the expertise of a professional forecaster. Cal Poly’s EE department has been collaborating with the Cal Poly Strawberry Center to investigate the potential in using digital imaging processing to predict harvests more reliably. This paper shows the progress of that ongoing project, as well as what aspects could still be improved. Three main blocks comprise this system: data acquisition, which obtains and catalogues images of the strawberry plants; computer vision, …

Developing, Evaluating, And Demonstrating An Open Source Gateway And Mobile Application For The Smartfarm Decision Support System, Caleb D. Fink

Master's Theses

The purpose of this research is to design, develop, evaluate, and demonstrate an open source gateway and mobile application for the SmartFarm open source decision support system to improve agricultural stewardship, environmental conservation, and provide farmers with a system that they own. There are very limited options for an open source gateway for collecting data on the farm. The options available are: expensive, require professional maintenance, are not portable between systems, improvements are made only by the manufacturer, limited in customization options, difficult to operate, and data is owned by the company rather than the farmer. The gateway is designed …