A Spin-off Version of Jason for IoT and Embedded Multi-Agent Systems

Carlos Eduardo Pantoja, Vinicius Souza de Jesus, Nilson Mori Lazarin, José Viterbo

Abstract: Embedded artificial intelligence in IoT devices is presented as an option to reduce connectivity dependence, allowing decision-making directly at the edge computing layer. The Multi-agent Systems (MAS) embedded into IoT devices enables, in addition to the ability to perceive and act in the environment, new characteristics like pro-activity, deliberation, and collaboration capabilities to these devices. A few new frameworks and extensions enable the construction of agent-based IoT devices. However, no framework allows constructing them with hardware control, adaptability, and fault tolerance, besides agents’ communicability and mobility. This work presents an extension of the Jason framework for developing Embedded MAS with BDI agents capable of controlling hardware, communicating, and moving between IoT devices capable of dealing with fault tolerance. A case study of an IoT solution with a smart home, a monitoring center, and an autonomous vehicle is presented to demonstrate the framework’s applicability.