No description available.

Substantial interest resides in identifying sensors, algorithms and fusion theories to detect explosive hazards. This is a significant research effort because it impacts the safety and lives of civilians and soldiers alike. However, a challenging aspect of this field is we are not in conflict with the threats (objects) per se. Instead, we are dealing with people and their changing strategies and preferred method of delivery. Herein, I investigate one method of threat delivery, side attack explosive ballistics (SAEB). In particular, I explore a vehicle-mounted synthetic aperture acoustic (SAA) platform. First, a wide band SAA signal is decomposed into a higher spectral resolution signal. Next, different multi/hyperspectral signal processing techniques are explored for manual band analysis and selection. Last, a convolutional neural network (CNN) is used for filter (e.g., enhancement and/or feature) learning and classification relative to the full signal versus different subbands. Performance is assessed in the context of receiver operating characteristic (ROC) curves on data from a U.S. Army test site that contains multiple target and clutter types, levels of concealment and times of day. Preliminary results indicate that a machine learned CNN solution can achieve better performance than our previously established human engineered Fraz feature with kernel support vector machine classification.

No description available.

Roadway construction workers have to work in close proximity to construction equipment as well as high-speed traffic, exposing them to an elevated risk of collisions. This research aims to develop an innovative holistic solution to reduce the risk of collisions at roadway work zones. To this end, a connected hazard detection and prevention system is developed to detect potential unsafe proximities in highway work zones and provide warning and instructions of imminent threats. This connected system collects real-time information from all the actors inside and outside of the work zone and communicates it with a cloud server. A hazard detection algorithm is developed to identify potential proximity hazards between workers and connected/automated vehicles (CAV) and/or construction equipment. Detected imminent threats are communicated to in-danger workers and/or drivers. The trajectories and safety status of each actor is visualized on Virginia Connected Corridors (VCC) Monitor, a custom web-based situational awareness tool, in real-time. To assure the accuracy of hazard detection, the algorithm accommodates various parameters including variant threat zones for workers-on-foot, vehicles, and equipment, the direction of movement, workers' distance to the work zone border, shape of road, etc. The designed system is developed and evaluated through various experiments on the Virginia's Smart Roads located at Virginia Tech. Data regarding activities of workers-on-foot was collected during experiments and was used and classified for activity recognition using supervised machine learning methods. A demonstration was held to evaluate the usability of the developed system, and the results proved the efficacy of the algorithm in successfully detecting potential collisions and provide prompt warnings and instructions. The developed holistic system elevates safety of highway construction and maintenance workers at work sites. It also helps managers and inspectors to keep track of the real-time safety status of their work zone actors as well as the accidents occurrences. As such, with the connected work zone hazard detection system, the safety level and productivity of the workers is expected to be greatly enhanced. / Master of Science / In order to reduce the risk of collisions for roadway construction workers, this research aims to develop an innovative holistic solution at roadway work zones. In this research, a connected hazard detection and prevention system is developed to detect potential collision hazards in highway work zones and generate warning and instructions of imminent threats. This system collects real-time information from all the workers, construction equipment and connected/automated vehicles (CAV) of the work. A hazard detection algorithm is developed to identify potential proximity hazards between them as well as to recognize the activities of workers. The trajectories and safety status of each worker, equipment or vehicle is visualized on Virginia Connected Corridors (VCC) Monitor, a custom web-based tool, in real-time. A demonstration was held to evaluate the developed system, and the results proved the efficacy of the algorithm in successfully detecting potential collisions and provide prompt warnings and instructions. The developed holistic system helps managers and inspectors to keep track of the real-time safety status of their work zone worker, equipment and vehicles as well as the accidents occurrences. As such, with the connected work zone hazard detection system, the safety level and productivity of the workers is expected to be greatly enhanced.

Different signal processing techniques for synthetic aperture acoustic (SAA) and highresolution voxel radar (HRVR) sensing modalities for side-attack explosive ballistic (SAEB) detection are proposed in this thesis. The sensing modalities were vehicle mounted and the data used was collected at an army test site. More specifically, the use of a frequency azimuthal (fraz) feature for SAA and the fusion of a matched filter (MF) and size contrast filter (SCF) for HRVR was explored. For SAA, the focus was to find a signature in the target’s response that would vary as the vehicle’s view on the target changed. For the HRVR, the focus was put on finding objects that were both anomalous (SCF) and target-like (MF). The results in both cases are obtained using receiver operating characteristic (ROC) curves and both are very encouraging.

In this dissertation, image feature extraction quality is enhanced through the introduction of two feature learning techniques and, subsequently, feature-level fusion strategies are presented that improve classification performance. Two image/signal processing techniques are defined for pre-conditioning image data such that the discriminatory information is highlighted for improved feature extraction. The first approach, improved Evolution-COnstructed features, employs a modified genetic algorithm to learn a series of image transforms, specific to a given feature descriptor, for enhanced feature extraction. The second method, Genetic prOgramming Optimal Feature Descriptor (GOOFeD), is a genetic programming-based approach to learning the transformations of the data for feature extraction. GOOFeD offers a very rich and expressive solution space due to is ability to represent highly complex compositions of image transforms through binary, unary, and/or the combination of the two, operators. Regardless of the two techniques employed, the goal of each is to learn a composition of image transforms from training data to present a given feature descriptor with the best opportunity to extract its information for the application at hand. Next, feature-level fusion via multiple kernel learning (MKL) is utilized to better combine the features extracted and, ultimately, improve classification accuracy performance. MKL is advanced through the introduction of six new indices for kernel weight assignment. Five of the indices are measured directly from the kernel matrix proximity values, making them highly efficient to compute. The calculation of the sixth index is performed explicitly on distributions in the reproducing kernel Hilbert space. The proposed techniques are applied to an automatic buried explosive hazard detection application and significant results are achieved.

Distracted driving involving secondary tasks is known to lead to an increased likelihood of being involved in motor vehicle crashes. Some secondary tasks are unnecessary and should never be performed. But other secondary tasks, e.g., operating the defroster, are critical to safe driving. Ideally, the driver should schedule when to perform the critical tasks such that the likelihood of a hazard materializing is relatively small during the performance of the secondary task. The current study evaluates a training program -- STRAP (Secondary Task Regulatory & Anticipatory Program) -- which is designed to make drivers aware of latent hazards in the hope that they regulate engagement in secondary tasks which they are performing at the time the latent hazard appears. The secondary tasks include both tasks that require drivers to take their eyes off the road (e.g., operating the defroster) and those which do not (e.g., cell phone use). Participants were assigned either to STRAP or placebo training. After training, the groups navigated eight different scenarios on a driving simulator and were instructed to engage during the drive in as many secondary tasks as possible as long as they felt safe to do so. Secondary task engagement was fully user paced. It is important to note that drivers receiving STRAP training were never instructed directly to either disengage from or not engage in secondary tasks when encountering latent hazards. The results show that STRAP trained drivers were more likely to detect latent hazards and associated clues than placebo trained drivers. With regards to secondary task engagement, STRAP trained drivers chose to limit their in-vehicle and cell phone task engagement by focusing on the forward roadway rather than the task at hand. STRAP training holds out the promise of providing individuals with the necessary skills and proactive awareness to make safe decisions regarding the non-performance or interruption of a secondary task in the presence of a potential latent hazard.

Two flash lidars, integrated from a number of cutting-edge components from industry and NASA, are lab characterized and flight tested under the Autonomous Landing and Hazard Avoidance (ALHAT) project (in its fourth development and field test cycle) which is seeking to develop a guidance, navigation, and control (GNC) and sensing system based on lidar technology capable of enabling safe, precise human-crewed or robotic landings in challenging terrain on planetary bodies under any ambient lighting conditions. The flash lidars incorporate pioneering 3-D imaging cameras based on Indium-Gallium-Arsenide Avalanche Photo Diode (InGaAs APD) and novel micro-electronic technology for a 128 x 128 pixel array operating at 30 Hz, high pulse-energy 1.06 ?m Nd:YAG lasers, and high performance transmitter and receiver fixed and zoom optics. The two flash lidars are characterized on the NASA-Langley Research Center (LaRC) Sensor Test Range, integrated with other portions of the ALHAT GNC system from around the country into an instrument pod at NASA-JPL, integrated onto an Erickson Aircrane Helicopter at NASA-Dryden, and flight tested at the Edwards AFB Rogers dry lakebed over a field of human-made geometric hazards. Results show that the maximum operational range goal of 1000m is met and exceeded up to a value of 1200m, that the range precision goal of 8 cm is marginally met, and that the transmitter zoom optics divergence needs to be extended another eight degrees to meet the zoom goal 6° to 24°. Several hazards are imaged at medium ranges to provide three-dimensional Digital Elevation Map (DEM) information. / Master of Science

<p dir="ltr"><a href="" target="_blank">ABSTRACT</a></p><p dir="ltr">A literature review was conducted, to determine and gain a better understanding of the environmental, technological, physiological, and psychological issues that lone agricultural workers potentially face in the event they are involved in an emergency. An investigation was conducted of communication devices used in other industries where working alone was common, to monitor for and detect incident occurrences. An assessment of currently available emergency alert software and sensing technology for <a href="" target="_blank">communication </a><a href="#_msocom_1" target="_blank">[AE1]</a> devices was also undertaken in this review.</p><p dir="ltr">Three hundred and sixty-eight U.S. cases of fatalities or injuries were analyzed in which working alone was identified as a contributing factor. Cases included lone agricultural workers, between the ages 15-64, who were identified from a convenient sample of incident reports from <a href="" target="_blank">2016-2021</a><a href="#_msocom_2" target="_blank">[AE2]</a> . Of the 368 lone agricultural worker incidents analyzed, 38% (140) were caused by tractor rollover or tractor runover, and ATV/ UTV rollovers. Grain bin entrapments accounted for 13% (48) of all identified cases, of which 86% (42) were fatal. Thirty-three percent (121) of the identified incidents involved equipment roll over (not including runovers), and 50% of identified victims, when age was known, were 57 years of age or older. In 11 cases (3%), the victim was under 15 years old and active in agricultural-related tasks at the time of incident occurrence.</p><p dir="ltr">Geospatial Information Systems (GIS) tools were used to identify the proximity of Emergency Medical Service (EMS) facilities and cellular towers from a convenient sample of 29 fatal and serious agricultural related injuries from 2016-2021, occurring in the state of Indiana. This analysis found that there were substantially fewer EMS facilities within close proximity to documented rural incident locations compared to injuries or fatalities occurring closer to a populated area. There were also fewer cellular towers within close proximity of incidents located primarily on or near rural agricultural land. More densely populated areas tended to have a greater density of EMS and cellular tower locations, with, most likely, more favorable outcomes from injuries due to shorter <a href="" target="_blank">response times.</a><a href="#_msocom_3" target="_blank">[AE3]</a></p><p dir="ltr">An investigation of the physical and operational impact that agricultural equipment would have on the efficacy of commercially available wearable technologies was undertaken, to detect the potential injury-causing agricultural incident. Five experiments were conducted to test the feasibility of these selected wearable devices in detecting agricultural-related incidents with the potential of causing serious injuries. Only one <a href="" target="_blank">simulated agricultural incident</a> <a href="#_msocom_4" target="_blank">[AE4]</a> successfully triggered incident detection. <a href="" target="_blank">Incidents successfully triggered incident detection on one wearable device, the Garmin Vivoactive 4 smartwatch. </a><a href="#_msocom_5" target="_blank">[AE5]</a></p><p dir="ltr">Recommendations included greater emphasis on the hazards associated with lone workers assigned agricultural workplaces, development of new, evidence-based educational resources to incorporate in current prevention strategies directed at farmers, ranchers, and agricultural workers, enhanced supervision of young agricultural workers and compliance with existing child labor regulations, equipping lone workers with appropriate cellphones and/ or wearable technologies to be carried in their vehicles, agricultural equipment, or on their person, use of electronic surveillance or monitoring equipment, written policies and procedures that enhance awareness of worker locations and conditions on a regular basis, and adherence to existing federal and state workplace safety and health regulations related to lone workers.</p><p><br></p><p dir="ltr"><a href="#_msoanchor_1" target="_blank">[AE1]</a>Changed</p><p dir="ltr"><a href="#_msoanchor_2" target="_blank">[AE2]</a>Fixed</p><p dir="ltr"><a href="#_msoanchor_3" target="_blank">[AE3]</a>I’m not sure how to address that more rural people are dying, given the limited scope and criteria for selection of the incidents selected in this study.</p><p dir="ltr"><a href="#_msoanchor_4" target="_blank">[AE4]</a>Not sure if this Is the best way to say it. I may end up cutting this part. I’ll pair the abstract down to ~250 words. For whatever reason, I thought the intro chapter abstract needed to be longer for a dissertation.</p><p dir="ltr"><a href="#_msoanchor_5" target="_blank">[AE5]</a>Shortened this paragraph and removed unnecessary detail, for clarity.</p>