Comparative study of Automotive Sensor technologies used for Unmanned Driving

Abstract

Autonomous vehicles utilize a large amount of data from Machine Learning, Neural networks, Image recognition systems for building the techniques that can drive autonomously. Autonomous vehicles depend on sensors for measuring conditions of roads and for making decisions while driving, and safety depends on the consistency of these sensors. Autonomous vehicles are robotic systems that are not only capable of regulating their motion in response to the sensory data they have obtained, but are also capable of behaving intelligently (or flexibly) in their environment. Autonomous vehicles must have the ability to see the things around it in order to know if they need to drive, to stop and turn, and handle the unexpected situations they come across. Each and every sensor has its own types of strengths and weaknesses in terms of range, recognition and reliability. Moreover, each sensor has its own advantages as well as disadvantages. This paper discusses the features of sensors used in autonomous vehicles and compares different set of sensors. We have used a Kalman filter for the detection and tracking of the car. We have used different parameters to see how tracking quality is affected by the tracker and also adjust the tracking filter to specify a different motion.