Spelling suggestions: "subject:"event aidentification"" "subject:"event anidentification""
1 |
Identification and Characterization of Damaging Road EventsAltmann, Craig Tyler 12 June 2020 (has links)
In the field of vehicle durability, many individuals are focusing on methods for better replicating the durability a user will experience throughout the typical design lifespan of a vehicle (e.g., 100,000 miles). To estimate user durability a means of understand the types of damaging events and driving styles of uses must be understood. The difficulty with accurately estimating customer usage is, firstly, there is a large pool of possible roads for a user to drive along, for example, there are over 4 million miles of public roads in the United States, alone [1]. In addition, while measurements of these surfaces could be collected it would be impractical for two reasons, the first is the financial and extreme time burden this would take. Second, when collecting measurements of a road surface only the current state of a road surface can be measured, thus as a road deteriorates or is repaved the measurements collected would no longer be an accurate representation of the road. It should be mentioned that even, if all of the road surfaces were measured performing simulation and analysis of all of these road surfaces would be computationally intensive. Instead, it would be beneficial if select events that account for a significant portion of the damage a vehicle experiences can be identified. These damaging events could then be used in more complex vehicle simulation models and as input and validation of proving ground and laboratory durability testing. The objective of this research is to provide a means for improved estimation of vehicle durability, specifically a means for identifying, characterizing, and grouping unique separable damaging events from a road profile measurement.
In order to achieve this objective a measure that can be used to identify separate damaging events from a road profile is developed. This measure is defined as Localized Pseudo Damage (LPD), which identifies the amount of damage each individual road excitation makes to the total accumulated damage for a single load path in a vehicle system. LPD is defined as a damage density to minimize the effect of measurement spacing on the resulting metric. The developed LPD measure is causal in that the value of LPD at a location is not affected by any future locations. In addition, for a singular event (e.g., impulse or step) in the absences of other excitations, the LPD value at the singular event location is equivalent to the total pseudo damage divided by the step size at the location.
Once a measure of pseudo damage density is known at multiple locations along a road profile for multiple load paths of interest, then separable damaging events can be identified. To identify separable damaging events the activity of the vehicle system must be considered because separate damaging events can only occur when a region of inactivity is present across all load paths. Subsequently, an optimization problem is formed to determine the optimal active regions to maintain. The cost function associated with the optimization problem is defined to minimize the cost (number of locations maintained in damaging events) and maximize the benefit (the amount of pseudo damage maintained).
Lastly, a statistical test is developed to assess if separate damaging events can be considered to be from the same general class of events based on their damage characteristics. The developed assessment methods establish the similarity between two more separable damaging events based on application specific user defined inputs. In the development, two example similarity metrics are defined. The first similarity metric is in terms of distance and the second is in terms of likelihood (probability). The developed statistical analysis uses the current state-of-the-art in clustering algorithms to allow for multiple damaging events to be identified and grouped together. / Doctor of Philosophy / In the automotive field determining the level of damage a typical production vehicle experiences over its lifetime has always been a desirable criterion to identify. This criterion is commonly referred to as customer usage. By understanding the typical customer usage of a vehicle over the lifetime of a vehicle, automotive engineers are able to improve the design of vehicle components. The issue with defining customer usage is that there are millions of miles of roads that a customer can travel on and millions of customers that all have unique driving characteristics. While it is possible to collect measurements of these road surfaces to use in further vehicle simulations, it is not feasible both from a financial and time perspective. In addition, the simulation and analysis of all road surfaces would be computationally intensive. However, if select damaging events (regions of the road surface that excessively contribute to accumulated damage) are identified, then they can be used in more complex vehicle durability analyses with lower computational efforts.
In conventional damage analysis a total amount of accumulated damage is established for a known road surface. The issue with defining damage this way is that unique events which likely contributed a large amount of the accumulated damage cannot be identified. The first objective of this research is to define damage as a function of the vehicle's location along a road surface. Then, unique and separable damaging events can be identified and separated from sections of the road that do not significantly contribute to the accumulated damage. After defining this measure, an optimization problem is developed to identify damaging events based on maximizing the benefit (amount of damage accounted for in damaging events) and minimizing the cost (amount of road surface retained). Unique and separable damaging events are identified by solving this optimization problem.
While the optimization problem identifies unique, separable damaging events, it is likely that some damaging events contain similar characteristics to each other. When performing additional durability analysis, it would be beneficial to form connections between similar damaging events to allow for analysis to be performed based on groups of events. To identify damaging events with similar characteristics, a statistical analysis is developed as the last contribution of this work. By combining this analysis with current state-of-the-art clustering algorithms and user provided definitions based on applications, similar damaging events are able to be grouped together.
|
2 |
On-Board Data Processing and FilteringFaber, Marc 10 1900 (has links)
ITC/USA 2015 Conference Proceedings / The Fifty-First Annual International Telemetering Conference and Technical Exhibition / October 26-29, 2015 / Bally's Hotel & Convention Center, Las Vegas, NV / One of the requirements resulting from mounting pressure on flight test schedules is the reduction of time needed for data analysis, in pursuit of shorter test cycles. This requirement has ramifications such as the demand for record and processing of not just raw measurement data but also of data converted to engineering units in real time, as well as for an optimized use of the bandwidth available for telemetry downlink and ultimately for shortening the duration of procedures intended to disseminate pre-selected recorded data among different analysis groups on ground. A promising way to successfully address these needs consists in implementing more CPU-intelligence and processing power directly on the on-board flight test equipment. This provides the ability to process complex data in real time. For instance, data acquired at different hardware interfaces (which may be compliant with different standards) can be directly converted to more easy-to-handle engineering units. This leads to a faster extraction and analysis of the actual data contents of the on-board signals and busses. Another central goal is the efficient use of the available bandwidth for telemetry. Real-time data reduction via intelligent filtering is one approach to achieve this challenging objective. The data filtering process should be performed simultaneously on an all-data-capture recording and the user should be able to easily select the interesting data without building PCM formats on board nor to carry out decommutation on ground. This data selection should be as easy as possible for the user, and the on-board FTI devices should generate a seamless and transparent data transmission, making a quick data analysis viable. On-board data processing and filtering has the potential to become the future main path to handle the challenge of FTI data acquisition and analysis in a more comfortable and effective way.
|
3 |
A Semantic Situation Awareness Framework for Indoor Cyber-Physical SystemsDesai, Pratikkumar 29 May 2013 (has links)
No description available.
|
4 |
Urban Seismic Event Detection: A Non-Invasive Deep Learning ApproachParth Sagar Hasabnis (18424092) 23 April 2024 (has links)
<p dir="ltr">As cameras increasingly populate urban environments for surveillance, the threat of data breaches and losses escalates as well. The rapid advancements in generative Artificial Intelligence have greatly simplified the replication of individuals’ appearances from video footage. This capability poses a grave risk as malicious entities can exploit it for various nefarious purposes, including identity theft and tracking individuals’ daily activities to facilitate theft or burglary.</p><p dir="ltr">To reduce reliance on video surveillance systems, this study introduces Urban Seismic Event Detection (USED), a deep learning-based technique aimed at extracting information about urban seismic events. Our approach involves synthesizing training data through a small batch of manually labelled field data. Additionally, we explore the utilization of unlabeled field data in training through semi-supervised learning, with the implementation of a mean-teacher approach. We also introduce pre-processing and post-processing techniques tailored to seismic data. Subsequently, we evaluate the trained models using synthetic, real, and unlabeled data and compare the results with recent statistical methods. Finally, we discuss the insights gained and the limitations encountered in our approach, while also proposing potential avenues for future research.</p>
|
Page generated in 0.1318 seconds