SMIA ecosystem: SMIA Operator

The SMIA Operator is a specialized tool that features a graphical interface allowing human users to easily interact with deployed agents. This tool provides a structured view of the functional information (CSS model) of the assets and enables users to request tasks from associated SMIA agents without having to deal with the complexity of the required I4.0 Language (FIPA-SMIACL).

Note

The SMIA Operator software has been used in the Use case transport logistics, which has an associated visual resource available on Youtube.

Source Code Reference

The SMIA Operator represents a human worker and acts as a Digital Twin (DT). Its Graphical User Interface (GUI) has been developed using native web interface functionalities of SPADE.

The source code can be downloaded directly from the official GitHub repository for manually provisioned environments.

Link to SMIA Operator source code

SMIA Operator agent source code

See also

The API documentation for the SMIA Operator source code is also available at API documentation SMIA ecosystem.

Deployment Environment

The first step is the establishment of an appropriate deployment environment. A valid environment containing SMIA Operator can be easily generated using the tool provided in this documentation platform: SMIA Environment Builder. To achieve the desired result, in step 3 SMIA Operator must be selected. There are two primary types of environments for the SMIA Operator.

Local deployment

In a local environment, SMIA instances are executed locally.

  1. Download the tool from the provided GitHub link.

  2. Configure the smia_operator_starter.py script. The JID must be configured to connect to the same XMPP server to enable agent communication.

  3. Specify the path for the SMIA_Operator_article.aasx model within the starter script. Verify that the model is located in smia_archive/config/aas.

  4. Execute the Python launcher script.

Virtualized deployment

For virtualized environments, Docker Compose or Kubernetes can be used. The SMIA Operator requires a specific Docker image built upon the core provided by SMIA. This image is available in the SMIA Docker Hub.

When deploying via Docker Compose, centralize the container and service definitions within a docker-compose.yml file.

  • The JID configuration is managed via the AGENT_ID and AGENT_PSSWD environment variables.

  • Verify that the SMIA_Operator_article.aasx model is located in the aas/ directory.

Execution is triggered by invoking the following command within the directory containing the YAML file and the aas/ folder:

docker compose up

If using a Kubernetes cluster, assuming the required YAML files are located in a deploy/ folder, execute the following command:

kubectl apply -f deploy/

Once deployed, each container initiates its lifecycle by autonomously self-configuring according to its corresponding AAS model.

Using the SMIA Operator

Following the successful instantiation of the SMIA agents, their performance and behavior can be validated. The SMIA Operator autonomously orchestrates the underlying FIPA-SMIACL interactions.

Dashboard layout

The SMIA Operator dashboard is divided into three main sections, each related to a different phase of the interaction process:

  • External SMIAs loader section: this section offers the possibility to discover the SMIAs available within the deployment environment and to show in a clear way the information extracted from their analysis. Since it is necessary to know the functionalities of the assets in relation to the CSS model, for each SMIA identified its associated AAS model is analyzed, obtaining all its CSS elements.

  • Available capabilities and skills section: this section shows by means of a table all the information related to the CSS model obtained from the analysis of the available SMIAs, but without distinguishing from which of them it has been obtained. The table is organized in identified capabilities and offers the possibility to select one of them.

  • SMIA candidates section: this section lists the SMIA instances that match the selected capability and its constraints. It allows the user to decide which specific agent (or group of agents for negotiation) should execute the request.

SMIA Operator dashboard

Figure: SMIA Operator dashboard

Accessing the GUI

To access the SMIA Operator control panel, open a web browser and navigate to the interface (see Figure: SMIA Operator dashboard):

http://localhost:10000/smia_operator

Note

In virtualized environments, the IP address should be changed to that of the container. It should also be verified whether the container’s exposed port matches or if a different one has been defined.

Execution Workflow

The user interface is deliberately designed to abstract underlying technical details and prioritize operational simplicity. The interaction follows a clear operational workflow:

  1. Load CSS Information: When accessing the page for the first time, only the top section will have content, as the SMIAs have not yet been loaded. At the top of the GUI, use the LOAD button to execute an automatic search and analysis of available SMIA agents. The GUI will be updated and all the information obtained will appear in the capabilities and skills table with the detected CSS elements.

  2. Select Capability: Target a specific capability using the SELECT button in the corresponding table row. If any extra data needs to be added (e.g., if the capability has more than one skill, you will be asked if you want to select a specific one; or if there are capability constraints, you will have to determine their values), the system interactively prompts for them. The SMIA candidates table is then updated with compatible assets.

  3. Request Execution: In the SMIA candidates table, decide which agent (a specific one or several to negotiate) should perform the capability, and use the REQUEST button to trigger execution.

    • If a single SMIA instance is selected, it directly executes the capability.

    • If multiple agents are selected, they initiate an autonomous distributed negotiation protocol (via FIPA-SMIACL) to determine the most suitable agent for the task.

    • If the selected skill has input parameters, their values must be provided before the request can be made.

Execution results

When the capability has been requested, the necessary interactions with the selected SMIAs are performed. Once the full request process is completed, the execution information is displayed on a new page (see Figure: SMIA Operator execution results page). This page is divided into two sections:

  • General information on the capability execution request.

  • A timeline with information on each step performed by the SMIA Operator, including interpretation steps and negotiation between agents.

This representation brings traceability and operational transparency to the distributed industrial environment.

SMIA Operator execution results

Figure: SMIA Operator execution results page