Design and Implementation of a Blind Spot Detection and Monitoring System for The Kayoola Buses.

Abstract

Changing lanes or negotiating a turn in a congested area while having no information about the objects in the blind spot area can be dangerous. It is particularly hard for drivers of the largest vehicles to see everything around them but the consequences of missing an obstruction could be catastrophic. As buses operate on increasingly crowded roads, drivers need to help in eliminating blind spots and highlight potential collisions before they occur. In this project, a Sensor-Based Blindspot Detection System for the Kayoola Buses to detect objects was proposed. This concept is implemented using a Raspberry Pi and two kinds of sensors; The Raspberry gathers input parameters from the two types of sensors (an ultrasonic sensor for object detection along with how far they are from the bus as well as an accelerometer for detecting motion in the bus). In addition, camera feeds are also fed to the Raspberry to allow for object detection. The microcontroller has predefined thresholds to decide if the measured value is off the range. If so, the driver is first alerted on LCD and LED, then later when the distance between the ego vehicle and the target gets smaller, auditory feedback is initiated as well. With the above setup, a hybrid blind spot detection system was feasible. In-vehicle tests were however not conducted, all tests were alpha tests performed on the bench. The project findings have shown that the HC-SR04 ultrasonic sensors and the Raspbian cameras are not suitable for deployment. For a complete deployable system, there is a need for identifying superior sensors, cameras, and microcontrollers that can easily be integrated with the CAN communication protocol.