OpTopNET: A Learning Optimal Topology Synthesizer for Ad-hoc Robot Networks

Abstract

In this paper, we synthesize a machine-learning stacked ensemble model a vector of which predicts the optimal topology of a robot network. This problem is technically a multi-task classification problem. However, we divide it into a class of multi-class classification problems that can be more efficiently solved. For this purpose, we first compose an algorithm to create ground-truth topologies associated with various configurations of a robot network. This algorithm incorporates a complex collection of nonlinear optimality criteria that our learning model successfully manages to learn. Then, we propose a stacked ensemble model whose output is the topology prediction for the particular robot associated with it. Each stacked ensemble instance constitutes three low-level estimators whose outputs will be aggregated by a high-level boosting blender. The results of the simulations, applying our model to a network of 10 robots, represents over %80 accuracy in the prediction of optimal topologies corresponding to various configurations of this complex optimal topology learning problem.