Comparison of WLAN Probe and Light Sensor-Based Estimators of Bus Occupancy Using Live Deployment Data

Area of Science:

• Transportation engineering
• Sensor networks
• Data fusion

Background:

• Bus operators need accurate, automated, low-cost passenger counting.
• Existing light sensor systems suffer from accumulating errors.
• Wireless Local Area Network (WLAN) probe-based systems face challenges in filtering external devices and mapping device counts to passenger numbers.

Purpose of the Study:

• To compare the accuracy and limitations of light sensor and WLAN probe-based passenger counting systems.
• To propose a novel data fusion approach to improve the accuracy of bus occupancy estimation.
• To develop a scalable and automated method for real-time passenger counting.

Main Methods:

• Comparison of light sensor and WLAN probe-based systems using real-world deployment data.
• Development of a fusion algorithm combining light sensor and WLAN data.
• Utilizing different error characteristics of each system for mutual error compensation.
• Calibration-free approach for parameter optimization.

Main Results:

• Both light sensor and WLAN systems have distinct trade-offs in accuracy and deployment.
• The proposed fusion approach significantly improves the accuracy of occupancy estimation.
• The fusion method effectively compensates for individual system errors.
• The approach is scalable, automated, and does not require ground truth data for calibration.

Conclusions:

• A novel fusion of light sensor and WLAN systems offers a more accurate and robust solution for bus passenger counting.
• The proposed method overcomes limitations of individual sensor systems.
• This approach provides a scalable, automated, and cost-effective solution for transportation operators.