Skip to main content

Software OTA

The Cogniteam Cloud Platform supports over-the-air (OTA) updates of its edge agents using Docker containers. This provides a scalable and reliable mechanism to roll out updates across fleets of robots or IoT devices with minimal manual intervention.

Overview​

Each edge agent is capable of provisioning Docker containers. The update process involves periodically checking and pulling updated images from a container registry.

This mechanism enables:

  • Controlled version upgrades
  • Rollback to previous stable versions

Update Flow​

  1. Start-Up Version Check Upon startup, the agent compares its current configuration version to the latest available version.

  2. Image Pull & Deployment If an update is needed:

    • The updated Docker image is pulled
    • The running container is stopped
    • A new container is launched with the updated image

Configuration​

Configuration is a composition of components. Configuration and Components are versioned in the system.

Example: component.json​

This section provides a sample JSON file to configure and deploy a component using the Cogniteam Cloud system.

What is a JSON File?​

JSON (JavaScript Object Notation) is a lightweight, human-readable data format commonly used for configuration and data exchange. It consists of attribute–value pairs and supports arrays and nested structures.

Creating a Component JSON File​

Once your component is containerized and available in a Docker registry, you can define a JSON configuration file that instructs the Nimbus system how to run it.

Sample Configuration:​

{
  "name": "nimbus/hector",
  "type": "COMPONENT_TYPE_COMPONENT",
  "className": "Ros1Component",
  "description": "2D laser scan mapping using Hector mapping algorithm",
  "environment": {
    "dockerInfo": {
      "image": "hector1:latest",
      "commands": [
        "roslaunch",
        "hector_mapping",
        "hector_mapping.launch",
        "map_resolution:=${map_resolution}",
        "map_size:=${map_size}",
        "base_frame:=${base_frame}"
      ]
    }
  },
  "parameters": {
    "parameters": [
      {
        "name": "map_resolution",
        "limits": {
          "min": 0.001,
          "max": 2,
          "step": 0.01
        },
        "doubleValue": 0.1
      },
      {
        "name": "map_string",
        "limits": {
          "min": 0,
          "max": 8,
          "step": 1
        },
        "stringValue": "string-value"
      },
      {
        "name": "base_frame",
        "limits": {
          "min": 0,
          "max": 8,
          "step": 1
        },
        "stringValue": "base_link"
      }
    ]
  },
  "streams": {
    "inputStreams": [
      {
        "name": "scan",
        "type": "",
        "cloud": {
          "id": "123",
          "download_policy": {
            "min_download_interval": "10000"
          }
        },
        "ros_topic": {
          "topic": "/scan",
          "type": "Messages.sensor_msgs.LaserScan"
        }
      }
    ],
    "outputStreams": [
      {
        "name": "map",
        "type": "",
        "ros_topic": {
          "topic": "/map",
          "type": "Messages.nav_msgs.OccupancyGrid"
        }
      },
      {
        "name": "pose_stamped",
        "type": "",
        "ros_topic": {
          "topic": "/slam_out_pose",
          "type": "Messages.geometry_msgs.PoseStamped"
        }
      }
    ]
  },
  "ros": {
    "base_frame": "base_link",
    "rate": 10.0,
    "publishTfDevices": true
  }
}

Explanation of Key Sections​

  • name: Component identifier
  • type: Component classification (e.g., software component)
  • className: Class type used by the Cogniteam Cloud system
  • environment.dockerInfo.image: Docker image tag
  • commands: Startup command and arguments passed to the container
  • parameters: Tunable settings with type, range, and default values
  • streams: Input and output ROS topics and their Nimbus equivalents
  • ros: TF settings, including base frame and publish rate

Advanced Docker Configuration​

"dockerInfo": {
  "image": "cognimbus/hamster_driver",
  "commands": [],
  "privileged": true,
  "gpu": true
}

Device Access Configuration​

"requiredDevices": [
  {
    "name": "laser",
    "info": {
      "type": "USB_PORT_TYPE_SERIAL",
      "productId": "ea60",
      "vendorId": "10c4"
    },
    "mountAs": "/dev/ttyUSB0"
  }
]

Uploading the JSON File​

You can register your component using the Cogniteam Cloud Web Interface

  • Navigate to: Components > Add Component
  • Enter a component name
  • Upload the JSON file
  • Then you will be able to use this new component inside a configuration

Local ROS Component (Non-Docker)​

The Agent can also run ROS components directly on the host without Docker, provided ROS is installed.

Example JSON for Local Component​

"environment": {
  "rosLocalInfo": {
    "PackageName": "hector_mapping",
    "launchFile": "mapping_default.launch",
    "workspaceSetup": "/opt/ros/melodic/setup.bash",
    "arguments": {
      "map_resolution": "${map_resolution}",
      "map_resolution2": "${map_resolution2}"
    }
  }
}