|
Tarun Kumar Singh
Er. Deepak Mathur
Keywords:
Pedestrian detection, infrared sensor, ultrasonic sensor.
Abstract:
Pedestrian safety is a growing concern, especially in urban areas where pedestrian-related accidents are prevalent. Traditional detection systems often struggle with environmental challenges such as low visibility during nighttime, fog, or rain, leading to inaccurate detections and increased risk. This project addresses these challenges by developing an advanced pedestrian detection system using a combination of infrared and ultrasonic sensors, simulated entirely in MATLAB. The proposed system leverages the strengths of both sensor types: infrared sensors are effective in low-light conditions, while ultrasonic sensors are robust in detecting pedestrians in poor visibility environments, such as fog or rain. Using MATLAB’s extensive toolboxes for signal processing, data acquisition, and sensor fusion, a real-time pedestrian detection system is developed. The system integrates both sensors to enhance detection accuracy and reliability, reducing false positives and improving response times.To optimize the system, various environmental conditions were simulated in MATLAB, including low-light, foggy, and rainy conditions. The pedestrian detection algorithms were fine-tuned to operate efficiently in real-time, leveraging MATLAB’s parallel computing capabilities. A weighted sensor fusion algorithm was employed to combine the data from infrared and ultrasonic sensors, ensuring the best performance across diverse environments. The system was validated through extensive testing under different environmental conditions. Comparisons with single-sensor systems demonstrated that the combined infrared-ultrasonic system significantly improves detection rates while reducing false positives, making it a reliable and efficient solution for enhancing pedestrian safety.
|
|
International Journal of Recent Research and Review
ISSN: 2277-8322
Vol. XVIII, Issue 1
March 2025
|
PDF View
PUBLISHED
March 2025
ISSUE
Vol. XVIII, Issue 1
SECTION
Articles
|
