Pedestrian Traffic Light Control with Crosswalk FMCW Radar and Group Tracking Algorithm.

Area of Science:

• Traffic Engineering
• Sensor Technology
• Artificial Intelligence

Background:

• Modern lifestyles increase traffic congestion, necessitating advanced traffic flow management.
• Static traffic light intervals are inefficient, particularly at pedestrian crossings.
• Existing detection systems like cameras have limitations in low visibility and privacy.

Purpose of the Study:

• To propose a smart pedestrian traffic light triggering mechanism.
• To leverage Frequency-modulated continuous-wave (FMCW) radar for reliable pedestrian detection.
• To enhance traffic flow by dynamically adjusting traffic signals based on real-time pedestrian presence.

Main Methods:

• Utilized an FMCW radar for pedestrian detection and data acquisition.
• Processed radar data into a point cloud structure for object representation.
• Applied a modified group tracking (GTRACK) algorithm on an external processor for pedestrian detection and tracking.
• Tested a prototype system designed to replace conventional pedestrian call buttons.

Main Results:

• The FMCW radar system reliably detected pedestrians in various scenarios.
• The modified GTRACK algorithm accurately identified pedestrian clusters within the radar's detection range.
• The prototype successfully triggered traffic signal changes only when pedestrians were present at the zebra crossing area.
• The system demonstrated effectiveness in removing the need for manual call buttons.

Conclusions:

• The proposed smart pedestrian traffic light system offers an effective, privacy-preserving alternative to traditional methods.
• FMCW radar combined with advanced tracking algorithms provides a robust solution for dynamic traffic management.
• This technology can significantly improve traffic flow efficiency and pedestrian safety at crossings.