Robotics Commons^™

A Pretraining-Based Framework For On-Device Training Of Imu-Based Locomotion Mode Detection For Wearable Active Exoskeletons, Muhammad Tahir Khan

LSU Master's Theses

Active exoskeletons are being developed to support human movement in physically demanding industries such as construction. For these systems to work effectively, they must be able to correctly identify the user’s current activity. This process is known as locomotion mode detection and plays an important role in selecting the appropriate control parameters for exoskeletons. Many existing approaches use inertial measurement units (IMUs) to recognize these activities and have shown strong performance. However, most of these methods depend on large amounts of labeled data collected under specific conditions. As a result, they often do not perform well when applied to new …

Identification Of Thruster Faults In Underwater Vehicles By Using Custom Encodings In Spiking Neural Networks, Donovan Gegg

LSU Master's Theses

Autonomous Underwater Vehicles (AUVs) are untethered robotic platforms used for tasks such as seafloor mapping, infrastructure inspection, and environmental monitoring. Recent technological advances have produced smaller, more affordable platforms, broadening access to research teams and small companies alike. This miniaturization comes at the cost of them handling drawbacks associated with a more compact machine such as reduced battery capacity as well as limited processing and sensing capabilities. These constraints make small-sized marine vehicle’s reliability critical as they can cause malfunctions, making the loss of a vehicle more likely. Actuator faults are particularly consequential as unintended and unstable control in an …

Seeing What Matters: Safety-Critical Semantic Segmentation Via Transfer Learning On Construction Sites, Obiora J. Odugu

LSU Master's Theses

Autonomous robots are increasingly deployed on construction sites for tasks such as progress monitoring, inspection, and safety assessment. For these robots to operate effectively, they must perceive and interpret complex, dynamic environments populated by workers, machinery, and unstructured terrain. Achieving reliable perception depends on high performing semantic segmentation models trained on large volumes of annotated data—an expensive and logistically challenging requirement in construction due to privacy restrictions, variable site access, and slow digitalization. This research addresses the challenge of limited labeled data by investigating transfer learning as a label-efficient approach for construction-site segmentation. Specifically, it explores whether road construction imagery—abundant …

Reality Capture Technologies For Continuous Inline Quality Monitoring In Construction 3d Printing, Michael J. Martin Ii

LSU Master's Theses

Extrusion-based construction 3D printing (C3DP) is an innovative and emerging technology aiming at the modernization of the construction industry. It uses robots and computer control to reduce the need for human labor and to decrease construction time and cost. C3DP creates elements layer by layer without using formworks, and it is critical to detect layer defects and deformations to ensure that each layer is printed with the desired dimensions. These defects can lead to a variety of issues during or after the construction process, including the collapse of freshly printed elements. In order to detect such deformations and to enable …

Lrtransformer: Learn-Region Transformer For Object-Agnostic Point Cloud Segmentation, Dipesh Gyawali

LSU Master's Theses

3D point cloud segmentation segments the 3D point cloud data into different regions/instances depending on their features that have numerous applications in robotics, autonomous driving, digital twinning, augmented reality, etc. The majority of the existing point cloud segmentation methods depend on class labels to identify 3D objects in the surroundings. Our work focuses on segmenting point clouds into different regions/instances in an object-agnostic manner for any number of objects in the environment. Given the point cloud, our method can segment the entire scene into multiple instances without depending on object shape and size. We leverage the power of the self-attention …

Autonomous Shipwreck Detection & Mapping, William Ard

LSU Master's Theses

This thesis presents the development and testing of Bruce, a low-cost hybrid Remote Operated Vehicle (ROV) / Autonomous Underwater Vehicle (AUV) system for the optical survey of marine archaeological sites, as well as a novel sonar image augmentation strategy for semantic segmentation of shipwrecks. This approach takes side-scan sonar and bathymetry data collected using an EdgeTech 2205 AUV sensor integrated with an Harris Iver3, and generates augmented image data to be used for the semantic segmentation of shipwrecks. It is shown that, due to the feature enhancement capabilities of the proposed shipwreck detection strategy, correctly identified areas have a 15% …

A Machine Learning Approach To Robotic Additive Manufacturing Of Uv-Curable Polymers Using Direct Ink Writing, Luis A. Velazquez

LSU Master's Theses

This thesis presents the design and implementation of a robotic additive manufacturing system that uses ultraviolet (UV)-curable thermoset polymers. Its design considers future applications involving free-standing 3D printing by means of partial UV curing and the fabrication of samples that are reinforced with fillers or fibers to manufacture complex-shape objects.

The proposed setup integrates a custom-built extruder with a UR5e collaborative manipulator. The capabilities of the system were demonstrated using Anycubic resin formulations containing fumed silica (FS) at varying weight fractions from 2.8 to 8 wt%. To fully cure the specimens after fabrication, a UV chamber was used. Then, measurements …

Enabling The Human Perception Of A Working Camera In Web Conferences Via Its Movement, Anish Shrestha

LSU Master's Theses

In recent years, video conferencing has seen a significant increase in its usage due to the COVID-19 pandemic. When casting user’s video to other participants, the videoconference applications (e.g. Zoom, FaceTime, Skype, etc.) mainly leverage 1) webcam’s LED-light indicator, 2) user’s video feedback in the software and 3) the software’s video on/off icons to remind the user whether the camera is being used. However, these methods all impose the responsibility on the user itself to check the camera status, and there have been numerous cases reported when users expose their privacy inadvertently due to not realizing that their camera is …

Device Free Indoor Localization Of Human Target Using Wifi Fingerprinting, Prasanga Neupane

LSU Master's Theses

Indoor localization of human objects has many important applications nowadays. Proposed here is a new device free approach where all the transceiver devices are fixed in an indoor environment so that the human target doesn't need to carry any transceiver device with them. This work proposes radio-frequency fingerprinting for the localization of human targets which makes this even more convenient as radio-frequency wireless signals can be easily acquired using an existing wireless network in an indoor environment. This work explores different avenues for optimal and effective placement of transmitter devices for better localization. In this work, an experimental environment is …

Practical Considerations And Applications For Autonomous Robot Swarms, Rory Alan Hector

LSU Doctoral Dissertations

In recent years, the study of autonomous entities such as unmanned vehicles has begun to revolutionize both military and civilian devices. One important research focus of autonomous entities has been coordination problems for autonomous robot swarms. Traditionally, robot models are used for algorithms that account for the minimum specifications needed to operate the swarm. However, these theoretical models also gloss over important practical details. Some of these details, such as time, have been considered before (as epochs of execution). In this dissertation, we examine these details in the context of several problems and introduce new performance measures to capture practical …

Fatigue Monitoring Through Wearable Sensors For Construction Workers, Srikanth Sagar Bangaru

LSU Doctoral Dissertations

About 40% of the US construction workforce experiences high-level fatigue, which leads to poor judgment, increased risk of injuries, a decrease in productivity, and a lower quality of work. Excessive fatigue from working in unpleasant working conditions, long working hours, or heavy workloads can aggravate fatigue's adverse effects, leading to work-related musculoskeletal disorders (WMSDs) and productivity loss. Therefore, it is essential to monitor fatigue to reduce the adverse effects and preventing long-term health problems. However, since fatigue demonstrates itself in several complex processes, there is no single standard measurement method for fatigue detection. This research aims to develop a system …

Distance-Based Formation Control Using Decentralized Sensing With Infrared Photodiodes, Steven Williams

LSU Master's Theses

This study presents an onboard sensor system for determining the relative positions of mobile robots, which is used in decentralized distance-based formation controllers for multi-agent systems. This sensor system uses infrared photodiodes and LEDs; its effective use requires coordination between the emitting and detecting robots. A technique is introduced for calculating the relative positions based on photodiode readings, and an automated calibration system is designed for future maintenance. By measuring the relative positions of their neighbors, each robot is capable of running an onboard formation controller, which is independent of both a centralized controller and a global positioning-like system (e.g., …

Reinforcement Learning Approach For Inspect/Correct Tasks, Hoda Nasereddin

LSU Doctoral Dissertations

In this research, we focus on the application of reinforcement learning (RL) in automated agent tasks involving considerable target variability (i.e., characterized by stochastic distributions); in particular, learning of inspect/correct tasks. Examples include automated identification & correction of rivet failures in airplane maintenance procedures, and automated cleaning of surgical instruments in a hospital sterilization processing department. The location of defects and the corrective action to be taken for each varies from task episode. What needs to be learned are optimal stochastic strategies rather than optimization of any one single defect type and location. RL has been widely applied in robotics …

A Low-Cost Experimental Testbed For Multi-Agent System Coordination Control, Victor Fernandez-Kim

LSU Master's Theses

A multi-agent system can be defined as a coordinated network of mobile, physical agents that execute complex tasks beyond their individual capabilities. Observations of biological multi-agent systems in nature reveal that these ``super-organisms” accomplish large scale tasks by leveraging the inherent advantages of a coordinated group. With this in mind, such systems have the potential to positively impact a wide variety of engineering applications (e.g. surveillance, self-driving cars, and mobile sensor networks). The current state of research in the area of multi-agent systems is quickly evolving from the theoretical development of coordination control algorithms and their computer simulations to experimental …

Reinforcement Learning In Robotic Task Domains With Deictic Descriptor Representation, Harry Paul Moore

LSU Doctoral Dissertations

In the field of reinforcement learning, robot task learning in a specific environment with a Markov decision process backdrop has seen much success. But, extending these results to learning a task for an environment domain has not been as fruitful, even for advanced methodologies such as relational reinforcement learning. In our research into robot learning in environment domains, we utilize a form of deictic representation for the robot’s description of the task environment. However, the non-Markovian nature of the deictic representation leads to perceptual aliasing and conflicting actions, invalidating standard reinforcement learning algorithms. To circumvent this difficulty, several past research …