41530 research outputs found
Sort by
Balancing Privacy and Accuracy in IoT using Domain-Specific Features for Time Series Classification
Get PDFε-Differential Privacy (DP) has been popularly used for anonymizing data to protect sensitive information and for machine learning (ML) tasks. However, there is a trade-off in balancing privacy and achieving ML accuracy since ε-DP reduces the model’s accuracy for classification tasks. Moreover, not many studies have applied DP to time series from sensors and Internet-of-Things (IoT) devices. In this work, we try to achieve the accuracy of ML models trained with ε-DP data to be as close to the ML models trained with non-anonymized data for two different physiological time series. We propose to transform time series into domain-specific 2D (image) representations such as scalograms, recurrence plots (RP), and their joint representation as inputs for training classifiers. The advantages of using these image representations render our proposed approach secure by preventing data leaks since these image transformations are irreversible. These images allow us to apply state-of-the-art image classifiers to obtain accuracy comparable to classifiers trained on non-anonymized data by ex- ploiting the additional information such as textured patterns from these images. In order to achieve classifier performance with anonymized data close to non-anonymized data, it is important to identify the value of ε and the input feature. Experimental results demonstrate that the performance of the ML models with scalograms and RP was comparable to ML models trained on their non-anonymized versions. Motivated by the promising results, an end-to-end IoT ML edge-cloud architecture capable of detecting input drifts is designed that employs our technique to train ML models on ε-DP physiological data. Our classification approach ensures the privacy of individuals while processing and analyzing the data at the edge securely and efficiently
Programmable Vibration Table Retrofit with Eagle Medical
Get PDFThis project aimed to repair and upgrade a vibration table used for package testing for the sponsor, Eagle Medical - a medical packaging and sterilization company based in Paso Robles, California. The current vibration table is not programmable and lacks circuit diagrams, making it difficult to repair. In addition, it is not functional due to internal circuitry issues and power requirements. To address these issues, the team will implement several upgrades to improve the product and allow for better random cyclic vibration testing. A full assessment of the existing vibration table will be conducted to ensure the viability of these proposed modifications for a successful outcome. The document provides information on the project\u27s background, objectives, and project management, including the design process, project timeline, and Gantt chart. Additionally, the document will provide information regarding our morphology, conceptual designs/evaluation, and failure analysis. This document also includes the detailed design, manufacturing plans, testing plans, and testing data/analysis, and instructions for use. We followed a 5-phase design process typically employed in industry: product discovery, project planning, product definition, conceptual design, and product development. During the Winter Quarter, we went through the first 4 phases, and all of Spring Quarter was spent in product development. In product discovery, we identified the needs of Eagle Medical and investigated other vibration tables on the market. During project planning, we mapped out our key deadlines using a Gantt chart. In product definition, we created the scope of the project and began creating our engineering specifications. In conceptual design, we took those specifications and started evaluating possible concepts using morphology. During product development, we built our prototype and tested it against our engineering specifications. The key customer requirements were that the new vibration table vibrates according to given standards, outputs frequency data, has vibration schedules, uses microcontrollers, has easily accessible data, updated/reliable electronics, and has a human-machine interface. The customer additionally wanted a circuit diagram, an easily serviceable table, and asked that we salvaged as much of the old table as possible. Our test plan revealed key results. The aesthetic satisfaction scored a 10. The vibration table achieved our max vibration goal of 18 m/s^2. We showed that the vibration table does have accurate control of the rotations per minute. We also showed that the motor wire temperature did not change, indicating that our wires are properly specified. We showed that the power supply had very little variance, so our electronics should be safe from voltage spikes. We showed that it takes approximately 5.2 seconds to upload a vibration test. We also showed that the emergency stop works with 100% certainty. Finally, we performed GUI unit testing and all units passed
Safety First Means Safety Fits- An Awareness Campaign for Inclusive PPE Designed for Women
Get PDFThe construction industry has long overlooked the unique needs of women in terms of personal protective equipment (PPE), leading to discomfort and safety risks. This paper presents the Safety First Means Safety Fits social media awareness campaign, which aims to increase awareness about the importance of size-inclusive PPE for women and the progress being made in the industry. The campaign utilized social media, specifically Instagram, as a platform for dialogue and information distribution. The campaign involved the creation of an Instagram account (@safetyfits_CP) and the development of six posts addressing various aspects of women\u27s PPE. The posts were promoted through other social media platforms, attracting followers, and generating interactions. Analytics data from Instagram\u27s business account platform were analyzed to evaluate the campaign\u27s effectiveness. Findings showed that the initial post, which introduced the campaign and its goals, received the most interaction, reaching a wide audience and attracting new followers. The post that advertised the second annual VEST Hackathon received the least interaction, possibly due to the use of less targeted hashtags. Overall, the campaign successfully increased awareness among 86 followers and received impressions from 280 accounts, fostering discussion and informing women about existing solutions and potential solutions. This paper highlights the importance of addressing gender disparities in PPE and demonstrates the potential of social media as a powerful tool for creating awareness and promoting change within the construction industry
