Error characterization The proposed system

TELKOMNIKA Indoor Localization Syst These Figures show each LED, with and without ex directly in front of the environm which means thatthe localizat was no problem related to an the typical size of a LED. neglected. Figure 4a. Original camer external illumination. Similar LED and the environment LEDs with a color iden for the system see Figures4 completely detected since the verify consistency such as LE the captured information in ord In these experiments an external illumination which illuminations to power up each environments since the tests d

4.2. Error characterization

This section describe with direct measurements ar innovative localization system of 25 series on the real robo also characterized in these te light sources is evaluated. Table 1. Error characte W Table 1 presents the artificial light the mean error lo infer that the environment’s lig this scale of accuracy. This m both conditions as the centroid ISSN: 1693-6930 ystem based on Artificial Landmarks and …. Andry w the detection of the artificial landmark with colo external illumination. It is possible to confirm that e onment light sources the LEDs are correctly detect ation system is not affected by external light. Durin an erroneous identification of the SCL, since the . Therefore, detected clusters of pixels without era image with ar color between a nt light sources. Figure 4b. Pre-processed im color identification of the ima illumination and similar color and the environment lig dentical to the environment source lights are not ve 4a and 4b. Nevertheless, see Figure 4b the the proposed algorithm has several steps that ma LED size, distance between adjacent LEDs, valid c order to retrieve a reliable extraction of the landmar ts we show that the SCL can be placed on the cei ich is an advantage because we can use the powe ach SCL. This turns the system flexible and more p s did not produce any erroneous localization for the bes the localization accuracy achieved by this sys are compared in order to characterize the error m. In the experiments, 50observations were made bot’s pose. The influence of the illumination factor tests as the accuracy of the system with and wit terization with and without light sources fluorescen Position Error m OrientationError º With 0.0283 0.645 Without 0.0248 0.705 he error of the proposed system. Although in the r lower obtained was lower, the gap is so small tha light sources do not have an influence on this sys minor gap is caused by the size of each LED wh oid position of the LEDs is used to compute the pos dry Maykol G. Pinto 615 olor identification of even with the SCL cted and identified, ring the tests, there e algorithm defines ut proper size are d image based on mage with external lor between a LED light sources. very recommended e yellow LED was make it possible to color sequence of ark. eiling, even near to wer supply of these e practical in indoor he robot. system. The results or obtained by this e per series a total tor in the system is ithout environment ent lamp. the absence of the that it is possible to system, considering which is different in ose estimation. TELKOMNIKA Vol. 10, No. 4 616

4.3. Comparison with the Sta