KMITL

Permanent URI for this communityhttps://dspace.kmitl.ac.th/handle/123456789/1

Browse

Settings

Access type: UnknownSubject: search.filters.subject.data-driven

Search Results

Now showing 1 - 3 of 3
  • Loading...
    Thumbnail Image
    Some of the metrics are blocked by your 
    consent settings
    Item type:Item,
    Optimizing Product Quality Prediction in Smart Manufacturing Through Parameter Transfer Learning: A Case Study in Hard Disk Drive Manufacturing
    (2025-04-01)
    Kaitwanidvilai, Somyot
    ;
    Sittisombut, Chaiwat
    ;
    Huang, Yu
    ;
    Bom, Sthitie
    In recent years, the semiconductor industry has embraced advanced artificial intelligence (AI) techniques to facilitate intelligent manufacturing throughout their organizations, with particular emphasis on virtual metrology (VM) systems. Nonetheless, the practical application of data-driven virtual metrology for product quality inspection encounters notable hurdles, such as annotating inspections in highly dynamic industrial environments. This leads to complexities and significant expenses in data acquisition and VM model training. To address the challenges, we delved into transfer learning (TL). TL offers a valuable avenue for knowledge sharing and scaling AI models across various processes and factories. At the same time, research on transfer learning in VM systems remains limited. We propose a novel parameter transfer learning (PTL) architecture for VM systems and examine its application in industrial process automation. We implemented cross-factory and cross-recipe transfer learning to enhance VM performance and offer practical advice on adapting TL to meet individual needs and use cases. By leveraging extensive data from Seagate wafer factories, known for their large-scale and high-dimensional nature, we achieved significant PTL performance improvements across multiple performance metrics, with the true positive rate (TPR) increasing by 29% and false positive rate (FPR) decreasing by 43% in the cross-factory study. In contrast, in the cross-recipe study, TPR increased by 27.3% and FPR decreased by 6.5%. With our proposed PTL architecture and its performance achievements, insufficient data from the new manufacturing sites, new production lines and new products are addressed with shorter VM model training time and smaller computational power with strong final quality prediction confidence.
  • Loading...
    Thumbnail Image
    Some of the metrics are blocked by your 
    consent settings
    Item type:Item,
    Comparative Foot Traffic Analysis During Normal Periods and Firework Events Using Wi-Fi Sensors
    (2024-01-01)
    Traganmaturapot, Peerada
    ;
    Sonehara, Noboru
    ;
    Hiruma, Nobuharu
    ;
    Cooharojananone, Nagul
    ;
    Kodate, Akihisa
    Recent observations indicate an increase in the frequency of crowd crush incidents, highlighting the urgent need for effective mitigation strategies. Addressing this issue necessitates a comprehensive understanding of the factors influencing visitor decision-making to prevent adverse outcomes such as crowd crushes. This study introduces a Streamlit dashboard designed to visualize foot traffic data in the Sendagaya area and integrates contextual data from 10 key factors, including events and locations, points of interest (POIs), periods of time, online search activity, height of buildings, temperature and weather conditions, currency exchange rates, earthquakes, number of international flight arrivals, and hotel room rates. This integration facilitates comparative analysis of foot traffic patterns during standard periods versus periods coinciding with significant events, such as the 2023 Jingu Gaien Fireworks Festival, to assess their impact on congestion levels. Conducted exclusively in Sendagaya, the study utilized 8 Wi-Fi sensors throughout August 2023, encompassing 3 key stages: data collection, preprocessing, and dashboard development. The analysis revealed significant determinants-including events, points of interest, time periods, and online activity-that influence visitor foot traffic, while other factors exhibited no discernible impact. These findings have important implications for enhancing decision-making processes, preparedness measures, risk management strategies, and data-driven policymaking for sustainable tourism development.
  • Loading...
    Thumbnail Image
    Some of the metrics are blocked by your 
    consent settings
    Item type:Item,
    Foot Traffic Analysis Using Wi-Fi Sensor During the Tokyo 2020 Olympics and Paralympics
    (2023-01-01)
    Traganmaturapot, Peerada
    ;
    Sonehara, Noboru
    ;
    Hiruma, Nobuharu
    ;
    Cooharojananone, Nagul
    ;
    Jirapongwanich, Jirakit
    Various real-world factors, such as time, weather, distance, environment, the COVID-19 pandemic, or even protests, can all impact human decision-making. However, restrictions and unexpected occurrences may also influence people's decisions regarding their path at any given time. These factors can lead to challenges in managing foot traffic at largescale events. In response to these challenges, this paper proposes a data-driven web-based foot traffic management supporting dashboard for large-scale events based on limited pedestrian count data, consisting of sensor name, latitude, longitude, MAC address, Datetime, and RSSI, collected by Wi-Fi sensors around the Sendagaya area during the Tokyo 2020 Olympics and Paralympics. The results confirmed that our proposed web-based dashboard contributes to human behavior understanding and decision-supporting policymaking for foot traffic management, which improves the design of spectator movement between transportation and venues in large-scale events. Furthermore, the dashboard is valuable from various perspectives, including preventing crowd crushing, redesigning areas to increase engagement in the shopping district, and improving traffic management.