Pedestrian Detection System with Speed Control

International Journal of Research and Scientific Innovation (IJRSI) | Volume V, Issue VI, June 2018 | ISSN 2321–2705

Pedestrian Detection System with Speed Control

Shashikala R^#, Chaitra^#, Kavya^#, Krithika Pai^#, Chaitra^#

^# Electronics and Communication Department, Shri Madhwa Vadiraja Institute of Technology and Management, Udupi, Karnataka, India.

Abstract— Pedestrian are the most vulnerable users of the road. There are millions of vehicles on the road and the conditions such as inclement weather, poor lightning, traffic and other road hazards restrict the visibility of drivers which increases the risk to pedestrian. In addition, human error is known to be one of the leading causes of accidents. This paper focuses on detection of pedestrian using real time video stream. Video is captured using Raspberry Pi camera module. Along with this Haar cascade classifiers are used as a technique to detect pedestrian in the video captured. An additional security is provided using obstacle detection implemented using an Ultrasonic sensor and speed control using PWM technique based on the position of the obstacle detected.

Keywords— Haar cascade classifier, Ultrasonic sensor, PWM, Raspberry Pi.

I. INTRODUCTION

In recent years, many auto manufacturers along with other technology based companies have done extensive research on pedestrian detection systems, integrating output signals into the control system of cars, as away to further curb pedestrian fatalities. Most pedestrian detection systems use sensors, which are integrated into the vehicle or attached with an after-market kit, in order to ‘sense’ the pedestrians in the nearby vicinity of the vehicle and alert the driver with warnings. Recent research in the field of Intelligent Vehicles has focused on using advances in information technologies to prevent human error. Advanced Driver Assistance Systems(ADAS)try to warn and prepare the driver in the event of possibly hazardous situations. Among ADAS, pedestrian detection is an important issue due to the vulnerability of pedestrians in the event of accidents. Some modern built-in sensor systems are able to output warning signals directly to the car’s control system and apply automatic breaking to help avoid a crash. Modern pedestrian detection systems use a variety of techniques to ‘sense’ the presence of pedestrians, such as infrared sensors, visibility light cameras, laser and radar technologies. In order to process the data from these sensors microcontrollers and microprocessors are often used.

Read More