Autonomní robot řízený mobilním telefonem s OS Android

Abstract

The aim of my thesis was to make an Android application which will be able to drive a car autonomously. For this purpose, the application uses a Bluetooth technology, a lidar sensor ( which detects obstacles) and also an object detection in a real time. These technologies should be helpful for better detection of obstacles and should be used for better control while driving a car. I focused on controlling a car by a gps sensor and current car's position .I solved the problem with GPS points, direction and compass by Android libraries.I can detect various obstacles in real time by using OpenCV. In a real simulation the car was not able to drive itself autonomously. The problem began when manipulating the steering wheel because it could not have got back to the original position. It's very difficult to predict the car's direction. I have also found out that the GPS locator delays in the mobile device and the detection of objects is very slow in a real time (sometimes about 0,5 frames per second) due to the high memory consumption. It was tested on the mobile Google Galaxy Nexus. The finding of my thesis is that detection of objects in a real time and navigation of a car is possible but there has to be a high-performance PC and better car driving. Nevertheless, my thesis' results enable another researches of this topic and it is possible to improve the methods of driving an autonomous car or to use other technology than Android.