MADE allows real-time control of field devices with low latency, gathers diagnostic information, and enables monitoring through the forwarding of alarms and event management.

By connecting a limit switch sensor or any on/off sensor to one of the enabled inputs, it is possible to trigger an outgoing alarm, such as a siren, spotlight, or any device that can be activated by the control unit. The solution is based on a cloud architecture, different from equivalent legacy systems, eliminating the need for local virtual machines for the necessary application components. Instead, it utilizes a board capable of connecting directly to AWS cloud and sharing relevant information to issue alarms when necessary. Additionally, an automatic system is present to detect the accessibility of each control unit. If any control unit is not reachable, an alarm is propagated to AWS (and then notifies the application dashboard accessed by operators).

The Raspberry microcontroller, connected to the single board designed and provided by System Management, features 16 inputs (covering 4 options) and 4 output relays to control up to 4 output devices. The reachability of the control units from the cloud will be ensured through the disconnection services (and automatic data resend) internally managed by AWS Greengrass. In addition to the presence checks provided out of the box by AWS Greengrass, a custom pinging mechanism is implemented using an MQTT heartbeat protocol.

In the system, the management of control events received on a second MQTT queue has been set up. In the same Python script responsible for sending data from the sensors, the sending of a keepalive event every 60 seconds has been arranged. This event indicates the proper integration and connection of each control unit to the AWS cloud environment, along with some telemetry data of the device.

Alarm management on local device

The management of events and relative alarms takes place via a Python script executed on the single control unit, capable of interfacing with the board and verifying the change in status of each single sensor, as well as piloting any connected relay outputs.

Interfacing is done through the use of the GPIO library which provides a simple interface to the devices and sensors connected and controlled via the GPIO pins.

The individual sensors are activated as shown in the table below:

Nel momento in cui uno dei sensori cambia il suo stato, interviene un servizio di gestione degli interrupt che in Python sul dispositivo locale, compone ed invia i dati dell’allarme, sotto forma di telemetria in AWS Greengrass, che poi successivamente avvia una chiamata GET HTTP, per l’integrazione con Dashboard allarmi del cliente.

Cloud alarm management

Integration with alarm Dashboard already in use by the customer is foreseen. The integration is carried out through HTTP GETs, parameterized and different according to the control unit (therefore the installation point) and the sensor. For each installation point, in fact, a "base URL" is associated which is common to all the devices connected to the control unit with the addition of parameters that differ, based on the type of sensor, the context and the criticality of the alarm. All the data generated by the sensors are - also - stored in the cloud application databases. For further information or for information on projects and applications, write to us using the contact form and we will provide you with all the information and support.