Centralized Algorithms Based on Clustering with Self-tuning of Parameters for Cooperative Target Observation

Clustering on target positions is a class of centralized algorithms used to calculate the surveillance robots' displacements in Cooperative Target Observation (CTO) problem. This work proposes and evaluates Fuzzy C-means (FCM) Density-Based Spatial Applications with Noise (DBSCAN) K-means (DBSk) based self-tuning clustering for CTO problem compares its performances that K-means. Two random motion patterns are adopted targets: free space or grid. As contribution, allows identifying ranges configuration parameters which each algorithm shows highest average performance. first conclusion, challenging situation relative speed targets high, sensor range low, existing present substantial drop performance, FCM proposed outperforms others. Finally, DBSk adapts very well low execution frequency, showing promising results this situation.