Cooperative Multi-agent on Head Yaw of Humanoid Robot Using Consensus

This paper presents the comparison between 2-proposed graph topologies. In this concern, each proposed graph topology consists of 3-nodes that represent a number of humanoid robots and 3-edges representing information exchange among robots. Two proposed graph topologies have been implemented on head yaw of humanoid robots. From the experimental results, two proposed graph topologies have been successfully applied to reach consensus. But, each proposed graph topology has different characteristic. The cyclic graph topology is similar to unsupervised learning characteristic because this topology cannot ensure where the consensus value will be achieved. Meanwhile, non-cyclic graph topology has the same characteristic as supervised learning where consensus value can be reached based on the predefined target value.