Analysis of Obstacle Detection Using Ultrasonic Sensor

Christofer N. Yalung,Cid Mathew Adolfo

Abstract:1Military Technological College, Muscat, Oman 2Military Technological College, Muscat, Oman ---------------------------------------------------------------------***--------------------------------------------------------------------Abstract – Automation of the car braking system is an important feature in the development of the smart car. The ability of a smart car to detect and classify an obstruction that is in varying proximities from it play a vital role in the system’s design. In this study, EV3 Lego Mindstorm equipped with an ultrasonic sensor was used as a model of a large scale vehicle. EV3 Lego MIndstorm was programmed to slow down when it is at a certain distance from the obstruction, and to stop when it is 15 cm away from the obstruction. There were five obstructions: wood, paper, cloth, plastic and metal. The distance measurement of the ultrasonic sensor and the Neural Network was used for the classification of the obstruction and Multiple Correlation was used for obstacle detection. There were 250 samples taken from the distance measurements of five different types of obstruction, each with a different cross sectional area, and a total recording time of 8 seconds. Overall, there is a high correlation coefficient in the distance measurement of the different types of obstruction materials. It is concluded that the ultrasonic sensor was able to detect the five given types of obstruction. Classification performance was very poor, which means that on the basis of distance measurement, the ultrasonic sensor cannot effectively classify the types of obstructions.

Physics,Engineering