METHODS OF LOCAL NAVIGATION OF MOBILE ROBOTS BASED ON OF TECHNICAL VISION
Abstract
The analysis of existing algorithms of local navigation is given, the possibility of applying the
existing algorithms of local navigation for mobile robots with the system of technical vision is investigated. A
complex algorithm of local navigation is proposed that allows solving the problems of local navigation in
conditions of incomplete awareness of the external environment.
In the work, the functioning of the localization navigation algorithm, Laturation, was carried out. For
this purpose, a wheeled robot was built on the platform Raberry Pi 2, equipped with a video camera as a
means of technical vision. Appearance of the developed robot is shown in Fig.4. The scheme of the experiment consisted in placing the robot in a medium with known visual reference points, after which the robot
carried out recognition of visual landmarks using the algorithm of Hahr's cascades, then calculated the state
to each landmark and formed a model of space. Then, a lateration algorithm was performed to determine the
coordinates. The robot software is implemented using the Python programming language. As a means of displaying information in the work, a screen is used. To carry out the reliability of the results, the distance to
the visual reference points was measured using conventional means of measurement, and then the location of
the robot camera was calculated.
As a result of the experimental study, the following results were obtained, the accuracy of determining
the distance to the landmarks was 8%, which is due to the accuracy of determining the camera's focal length.
The accuracy of determining the coordinates of the camera of the mobile robot was 6.5%, which is acceptable for the conditions of the experiment