ANALIZA PERFORMANSI ROBOTA POMOĆU FUZZY GRAFOVA I FUZZY C-MEANS ALGORITMA
PERFORMANCE ANALYSIS OF ROBOTS USING FUZZY GRAPHS AND THE FUZZY C-MEANS ALGORITHM

Summary/Abstract: This paper presents an approach to robot performance analysis that combines two complementary concepts: fuzzy graphs and the Fuzzy C-Means (FCM) algorithm. The goal of the paper is to demonstrate how robot accuracy and repeatability can be combined into a unified reliability measure, and how this measure can be further used to form a fuzzy similarity index between robots. When applied to a dataset of 16 simulated robots, the FCM algorithm forms four clearly separated clusters with an average membership degree of 0.97. The obtained membership values are then used as fuzzy weights for the construction of a fuzzy graph. The analysis of this graph shows that, for a similarity threshold of 𝜏 = 0.95, the structure changes from four to six clusters, while 87.5% of the robots remain connected within stable subgroups. This approach enables a deeper understanding of the relationships between robots—on one hand through graph structures with fuzzy weights, and on the other hand through grouping robots based on fuzzy cluster membership. Such a hybrid approach contributes to better insight into the performance structure and may have practical applications in robot maintenance and task allocation.

• Details
• Contents