Cloud based Real-time Monitoring and Control System (ongoing)

This project uses cloud platform (IBM Cloud) to conduct process monitoring and control, e.g., additive manufacturing process monitoring and control: .

The system’s hardware include: sensors (gyroscope, camera etc.), Arduino, and Photon. To build the system, there are two major steps, i.e., (1) enable the connection of sensors-Arduino-Photon-IBMCloud; (b) create data analysis models and algorithms to analyze the input sensors data, and give output of monitoring results and control signals.

1. Build the connection of sensors-Arduino-Photon-IBMCloud

The connection includes: (a) connect multiple sensors with Arduino; (b) communications of Arduino-Photon (c) communications of Photon-IBMCloud.

1.1 Connect multiple sensors with Arduino

1.2 Communications of Arduino-Photon

1.3 Communications of Photon-IBMCloud

• Set up IBM Watson IoT, and add device to the IoT platform. An instruction can be found here.
• Write codes to Photon to enable the connection. The connection is via MQTT. A simple example that sends and analyzes random number is as follows:
  • generate 0-100 random numbers and send the number to IBM IoT (client.connect and client.publish);
  • use Node-RED to simply analyze the random number, i.e., if the number is bigger than a specified threshold, then return “1”, otherwise, return “0”. The returned commands “1” or “0” is sent back to IBM Cloud. The Node-RED flows can be found here;
  • receive the “1” or “0” commands from IBM Cloud to Photon (client.subscribe), and blink the LED lights if the command is “1”.
• The codes are as follows:

#include "MQTT/MQTT.h"

char *IOT_CLIENT = "d:org_id:device_type:device_id";
char *IOT_HOST = "org_id.messaging.internetofthings.ibmcloud.com";
char *IOT_PASSWORD = "_device_token_here_";
char *IOT_PUBLISH = "iot-2/evt/randomnumber/fmt/json";
char *IOT_SUBSCRIBE = "iot-2/cmd/alert/fmt/json";
char *IOT_USERNAME = "use-token-auth";

int count = 1;
int led1 = D0;
int led2 = D7;

void callback( char* topic, byte* payload, unsigned int length ) {
  payload[length] = '\0';
  String s = String((char*)payload);
  int f = s.toInt();
  //Particle.publish("payload", String(s), PRIVATE);
  if (f == 1){
     digitalWrite(led1, HIGH);
     digitalWrite(led2, HIGH);
     delay(100);
     digitalWrite(led1, LOW);
     digitalWrite(led2, LOW);
  }
}

MQTT client(IOT_HOST, 1883, callback);

void setup() {
  pinMode(led1, OUTPUT);
  pinMode(led2, OUTPUT);
  client.connect(
    IOT_CLIENT,
    IOT_USERNAME,
    IOT_PASSWORD
  );
  //client.subscribe( IOT_SUBSCRIBE );
  if (client.subscribe(IOT_SUBSCRIBE)) {
      Particle.publish("subscribe", String("subscribe OK"), PRIVATE);
    } else {
      Particle.publish("subscribe", String("subscribe failed"), PRIVATE);
    }

  if( client.isConnected() ) {
      Particle.publish("status", String("Connected"), PRIVATE);
    } else {
      Particle.publish("status", String("Unconnected"), PRIVATE);
    }
}

void loop() {
    client.publish(IOT_PUBLISH, "{\"value\": " + String(random(100)) + " }"
    );
    //Particle.publish("counting", String(count), PRIVATE);
    count = count + 1;
    client.loop();
    delay( 1000 );
}

• Notes: If the connection to IBM IoT failed, try modifying the security settings of IBM IoT: in “SECURITY-Connection Security”, change the “security level” to “TLS Optional”. Also, the Node-RED application must be built, otherwise, no feedback command is generated.

2. Store and analyze data using IBM cloud

• A cloud database storing historical data must be built;
• Algorithms analyzing streaming data and historical data should be proposed based on the specific monitoring/control task. (Streaming Analytics in IBM Cloud, and Node-RED are potentially to be used.)

Acknowledgment

This project is supported by the National Science Foundation, RISE NSF#1646897.