Comparison of Analytical Method and Artificial Neural Network Method for Calculation of Optical Positioning

This paper compares two methods of passive determination of an object (camera) position, using the optical beacon. The LED light sources as a part of the optical beacon are placed at certain positions to create a specific geometric pattern. With the changes in mutual camera and beacon position, there are also changes in the pattern, which is determined in the camera-captured picture. Due to a change of the camera position against the beacon, there is also a change in the mutual position of the light sources in the image captured by the camera. The first method is based on analytical expression of the beacon range and azimuthal angle in dependence on mutual positions of the light sources in the picture. The second method uses an artificial neural network instead of analytical expression. A description of both methods together with comparison of experimental results are presented in this paper.