Engineering Approaches for Programming Agent-Based IoT Objects Using the Resource Management Architecture

The Internet of Things (IoT) allows the sharing of information among devices in a network. Hardware evolutions have enabled the employment of cognitive agents on top of such devices, which could help to adopt pro-active and autonomous IoT systems. Agents are autonomous entities from Artificial Intelligence capable of sensing (perceiving) the environment where they are situated. Then, with these captured perceptions, they can reason and act pro-actively. However, some agent approaches are created for a specific domain or application when dealing with embedded systems and hardware interfacing. In addition, the agent architecture can compromise the system’s performance because of the number of perceptions that agents can access. This paper presents three engineering approaches for creating IoT Objects using Embedded Multi-agent systems (MAS)—as cognitive systems at the edge of an IoT network—connecting, acting, and sharing information with a re-engineered IoT architecture based on the Sensor as a Service model. These engineering approaches use Belief-Desire-Intention (BDI) agents and the JaCaMo framework. In addition, it is expected to diversify the designers’ choice in applying embedded MAS in IoT systems. We also present a case study to validate the whole re-engineered architecture and the approaches. Moreover, some performance tests and comparisons are also presented. The study case shows that each approach is more or less suitable depending on the domain tackled. The performance tests show that the re-engineered IoT architecture is scalable and that there are some trade-offs in adopting one or another approach. The contributions of this paper are an architecture for sharing resources in an IoT network, the use of embedded MAS on top IoT Objects, and three engineering approaches considering agent and artifacts dimensions.

Deliberative Agents in Dynamic Environments, Using Jason and NARS

This work explore the possibility to combine the Jason reasoning cycle with a Non-Axiomatic Reasoning System (NARS) to develop multi-agent systems that are able to reason, deliberate and plan when information about plans to be executed and goals to be pursued is missing or incomplete. The contribution of this work is a method for BDI agents to create high-level plans using an AGI (Artificial General Intelligence) system based on non-axiomatic logic.

Deliberative Agents in Dynamic Environments, Using Jason and NARS

This work explore the possibility to combine the Jason reasoning cycle with a Non-Axiomatic Reasoning System (NARS) to develop multi-agent systems that are able to reason, deliberate and plan when information about plans to be executed and goals to be pursued is missing or incomplete. The contribution of this work is a method for BDI agents to create high-level plans using an AGI (Artificial General Intelligence) system based on non-axiomatic logic.

Agent in a Box: A Framework for Autonomous Mobile Robots with Beliefs, Desires, and Intentions

This paper provides the Agent in a Box for developing autonomous mobile robots using Belief-Desire-Intention (BDI) agents. This framework provides the means of connecting the agent reasoning system to the environment, using the Robot Operating System (ROS), in a way that is flexible to a variety of application domains which use different sensors and actuators. It also provides the needed customisation to the agent’s reasoner for ensuring that the agent’s behaviours are properly prioritised. Behaviours which are common to all mobile robots, such as for navigation and resource management, are provided. This allows developers for specific application domains to focus on domain-specific code. Agents implemented using this approach are rational, mission capable, safety conscious, fuel autonomous, and understandable. This method was used for demonstrating the capability of BDI agents to control robots for a variety of application domains. These included simple grid environments, a simulated autonomous car, and a prototype mail delivery robot. From these case studies, the approach was demonstrated as capable of controlling the robots in the application domains. It also reduced the development burden needed for applying the approach to a specific robot.