Model of trusted interaction of agents in decentralized cyber-physical environmentv

Abstract

Objective. The purpose of the work is to increase the efficiency of the functioning of agents of a cyber-physical system in the presence of agents with incorrect or malicious behavior. The goal is achieved by establishing a trusted interaction between agents and quick detection of malicious impact.Method. Trusted interaction is carried out using distributed ledger technology and agent confidence indicators. The novelty of the proposed solution lies in the fact that information about the actions of agents is stored and aggregated using smart contracts at specified time intervals. Each agent keeps a local copy of the agent interaction chain. If several agents interact with each other, then they exchange information stored in their copies of the ledger.Result. To test the proposed method, we implemented it in C++. For the experiments, the scenario of collective perception by agents of a decentralized cyber-physical environment in a specialized simulation program Contiki Cooja was used.Conclusion. The method implemented using the solutions proposed in this work showed higher efficiency than the method based on the dynamic calculation of the confidence index. The proposed solutions can be applied not only in cyber-physical systems, but also in any other systems with decentralized control.