Online System QoS Enhancement of Flipped Classroom for Intelligent Manufacturing Specialty Based on Intelligent Information Platform

Xianglin Wei, Xiaoli LiInternational Journal of Science and Engineering Applications 2023 · DOI

TL;DR

This paper proposes an end-to-end Quality of Service (QoS) enhancement framework for "flipped classrooms" within the Intelligent Manufacturing specialty. By leveraging an Intelligent Information Platform and optimizing network-layer protocols—specifically through packet prioritization and bandwidth allocation in Ethernet Passive Optical Networks (EPON)—the authors demonstrate how to maintain high-quality online instruction for complex, data-heavy subjects like machine vision.

What Problem Does This Paper Address?

The rapid shift to online learning, accelerated by the COVID-19 pandemic, exposed significant gaps in teaching technical, hands-on engineering disciplines. Specifically, courses like "Machine Vision and Inspection" require high-bandwidth, low-latency data transmission to support real-time software interaction and high-definition video streaming.

Traditional network setups often struggle with "best-effort" delivery, leading to jitter and packet loss that disrupt the "flipped classroom" model—where students engage with primary content independently and use "class time" for interactive problem-solving. The authors identify a need for a centralized management system that can guarantee service quality at the network level to ensure that students remain engaged and can complete complex exercises in real-time.

Methods

The authors propose a multi-layered approach centered on an Intelligent Information Platform. Key technical components include:

  • Network Architecture (EPON): The system utilizes Ethernet Passive Optical Network (EPON) technology. To manage the distance-related delays between the Optical Line Terminal (OLT) and various Optical Network Units (ONUs), the system uses time-stamping for automatic ranging and delay adjustment.
  • Bandwidth Allocation: The paper discusses the implementation of Dynamic Bandwidth Allocation (DBA) versus static methods to better handle the fluctuating data demands of a virtual classroom.
  • Packet Classification: A core feature is the ability for network managers to define six levels of service priority. By utilizing three priority bits in the packet header, the system ensures that critical teaching data (like live video or control signals for machine vision software) is prioritized over less sensitive background traffic.
  • Software Integration: The curriculum is built around the LabVIEW platform and Visual Assistant 2014, allowing students to perform image processing operations (histogram analysis, geometric transformations, etc.) remotely while the platform monitors the end-to-end QoS.

Datasets

The paper does not utilize a single, centralized public dataset for training a model. Instead, it references:

  • Pilot Program Data: Results from a pilot implementation within the authors' school network engineering major.
  • Sensory Data: The platform is designed to handle "massive sensory data" characterized by heterogeneous formats and large scales, drawing examples from Intelligent Transportation Systems (ITS) and power grid monitoring to prove the platform's robustness.

Key Insights and Ideas

  • The Data Link Layer is Critical: A major insight is that upper-layer protocols (like HTTP or standard streaming protocols) cannot solve service differentiation alone. True QoS must be addressed at the data link layer of the EPON to control delay and jitter effectively.
  • Subjective vs. Objective Quality: The authors emphasize that "Quality of Service" is ultimately defined by the end-user's subjective acceptability. Technical metrics (delay, loss) are only useful insofar as they improve the student's actual experience.
  • The "Splicing" Challenge: The paper highlights a common issue in vocational education where practical training is "spliced" onto theoretical courses haphazardly. An integrated intelligent platform helps bridge this gap by providing a unified environment for both theory and practice.
  • Versatility of the Platform: While focused on education, the authors illustrate the platform's power by showing its application in monitoring illegal vehicles via license plate recognition and detecting faults in power supply equipment.

Key Results

The pilot results indicate that the characteristic reform of network engineering-related majors can effectively meet local demands for technical talent. Qualitatively, the authors found that there is "not much difference" between the effectiveness of online and offline teaching for machine vision, provided the underlying platform supports:

  • High Accuracy and Efficiency: Through proper load balancing and QoS monitoring.
  • Real-time Interaction: Allowing students to watch replays (e.g., via Tencent Classroom) and interact with instructors without technical lag.
  • Centralized Management: Combining deployment and monitoring into a single end-to-end solution.

Limitations and Future Work

The authors acknowledge that while the platform is robust, certain applications—such as freight and logistics information security—still suffer from low accuracy and security, leading to potential losses.

Future work will focus on:

  • Quantifying Service Quality: Developing a more rigorous index system to guide the operation and maintenance of the network.
  • Talent Cultivation: Addressing the "huge demand" for engineers capable of constructing ubiquitous networks for industry and commerce over the next decade.
  • Refining Load Balancing: Improving the technology to ensure the overall efficiency of the flipped classroom remains stable as student numbers scale.

Summary updated

Abstract

In-depth analysis of the online teaching mode of the "Machine Vision and Inspection" course of the intelligent manufacturing major, and research on the teaching work of the machine vision and inspection course from the teaching objectives and positioning and the key content of the course. And in our school network engineering major Corresponding experiments were carried out. The pilot results show that, for ordinary colleges and universities aiming at cultivating applied talents, the characteristic reform of the original network engineering-related majors can effectively meet the local demand for talents in this field. It is an end-to-end QoS monitoring solution that combines deployment and centralized management.
BibTeX
@article{wei2023online,
  author = {Xianglin Wei and Xiaoli Li},
  title = {{Online System QoS Enhancement of Flipped Classroom for Intelligent Manufacturing Specialty Based on Intelligent Information Platform}},
  journal = {International Journal of Science and Engineering Applications},
  year = {2023},
  doi = {10.7753/ijsea1203.1028},
}
Figure