Technical Vision Based Autonomous Navigation Intelligent System of a Mobile Robot

Abstract

Intelligent autonomous navigation system based on technical vision is considered as part of a training and research project. The Autonomous navigation system and its elements are considered. Problems of creating a prototype are described. The goal of this project is to develop this system for a four-wheeled robot. There are some problems with design: the first one is the hard artificial intelligence (AI) part. The similar AI programms are made for autopilots by Tesla [1] and Google. Also, we are used to look at robots that can move straight. But if you want to turn right or left the robots have to stop and then swerve. This problem is called Dubin's path. The third problem is the problem of extracting unnecessary elements. They can either disorient the system or add extra calculations inside the program. Next problem is the problem of detecting human. We know that road safety is the most important part of transportation. You need to devote most of your time to this aspect. More precisely, the development of the system as a safety assistant.