People with lower-limb prostheses experience limited tactile and perceptual knowledge of their prosthetic limbs. This has been shown to contribute to improper gait kinematics, impaired balance, and musculoskeletal disorders. This work presents a novel haptic feedback system that aims to provide real-time augmented sensory feedback to people with lower-limb prostheses. The system consists of an insole with piezoresistive force sensors, a microcontroller, and a vibrotactile thigh sleeve with four pancake motors. Force data from the insole are used to calculate the plantar center of pressure, and changes in the center of pressure are then presented to the user as time-discrete vibrations on the medial thigh. Human perceptual testing was conducted to determine the efficacy of the proposed haptic display in conveying gait information to users. Thirteen able-bodied participants wearing the haptic sleeve were able to accurately identify differences in the speed of step patterns (92.3%) and to classify full or partial patterns (94.9%). These results suggest that the system was effective in communicating center of pressure information through vibrotactile feedback.

We rely on our hands in nearly all daily activities. This thesis focuses on activity tracking in small parts industrial assembly from a hand-centric viewpoint to enable more dynamic supervision in environments where workers frequently move within the assembly area. Additionally, the model is expected to be able to anticipate future actions and provide feedback for assistive assembly (Ragusa et al., 2023). Traditional action recognition models rely on an egocentric viewpoint such as models trained on the EgoHands dataset (Bambach et al., 2015) or the Assembly 101 model for small carts assembly activity recognition (Sener et al., 2022). These existing models struggle with precisely tracking and recognizing hand interactions within the workspace from hand-centric viewpoints in first-person industrial assembly settings. They often lack the ability to capture the dynamics between hands and objects, which is required for accurate small parts assembly tracking (Pei et al., 2025). This thesis proposes a computer vision-based system that detects and recognizes hands and assembly parts in the working area using YOLO (You Only Look Once) for object detection and tracking (Hashimoto et al., 2019), aiming to identify working hands and small parts of the assembly within the workspace.

A key challenge in computing education is gaining insight into a student's problem-solving process. Although problem-solving manifests in many computing skill areas, it is particularly prominent in programming. It has been suggested that neither programming homework, written exams, or oral exams are sufficient for measuring programming ability due to the process-centric nature of programming. Autograders are commonly introduced into programming courses to support assignment scalability and reproducibility. As a side effect of deploying an autograder in a course, a sequence of assessments for a programming assignment may be captured. This is a unique source of information about a student's problem-solving ability as seen through their programming activity. Developing methodologies to analyze this source of data provides insight into the process by which a student interacts with assignments. Student problem-solving is often understood by examining only the end result of a student's effort. This obscures many intermediate actions. For example, a student may have struggled with a part of an assignment, indicating under-preparedness or an inadequate assignment, but have finally stumbled onto the answer. Examining only the result means that many assessment methods provide only an indirect measure of a student's problem-solving ability, which is not ideal for instruction.This dissertation addresses these concerns by introducing a specialized automated assessment tool for intermediate computing courses and proposing three methods for analyzing student data. In the first method, a difficulty score is computed based on the pattern of test regressions during development. In the second, a graph representing how students progress through an assignment is constructed. In the third, a measure for how well a student adheres to a development process is calculated. Automatically analyzing the trace captured by an autograder has two major challenges: 1) the presence of commonsense reasoning that instructors informally apply to evaluate a student's problem-solving process, and 2) the combinatorial problems that emerge from the breadth of possible solutions.