A Measurement-Based Frame-Level Error Model for Evaluation of Industrial Wireless Sensor Networks

Area of Science:

• Computer Science
• Electrical Engineering
• Industrial Engineering

Background:

• Industrial wireless sensor networks (IWSNs) are crucial for smart manufacturing.
• Accurate wireless link characterization is needed for IWSN simulation and bottleneck identification before deployment.
• Existing IWSN simulators often lack error models based on real industrial environment data.

Purpose of the Study:

• To develop and validate a novel, data-driven error model for industrial wireless sensor networks.
• To improve the accuracy of IWSN performance simulations in manufacturing settings.
• To enhance the reliability estimation of IEEE 802.15.4 wireless links in industrial environments.

Main Methods:

• Conducted a one-day experiment in a manufacturing factory to measure IEEE 802.15.4 transmission quality.
• Constructed a second-order Markov frame-level error model using the collected measurement records.
• Integrated the proposed error model into the OpenWSN simulator.

Main Results:

• The proposed error model improved the accuracy of estimated transmission reliability by up to 12% compared to traditional models.
• Simulation accuracy showed improvement with increasing burst losses, a common characteristic of industrial wireless links.
• The model's effectiveness was demonstrated within the OpenWSN simulator, which implements relevant IEEE and IETF standards.

Conclusions:

• The developed Markov-based error model provides a more accurate representation of industrial wireless link behavior.
• This enhanced simulation capability aids in identifying performance bottlenecks and optimizing IWSN deployment for smart manufacturing.
• The study highlights the importance of using real-world data for developing effective IWSN simulation tools.