Smart Farm — Soil Moisture Based Automatic Irrigation

Overview

A smart farm is a system that applies sensors and communications (IoT), data analytics, and AI to agriculture (crops, livestock, and aquaculture) to automatically monitor and control the environment, thereby increasing productivity and quality while reducing water, energy, and labor. Among these, the automatic irrigation system is designed to keep soil moisture within a hysteresis-based target range, with minimum on-time and cooldown intervals enforced on the pump to prevent excessive switching and overheating. Measurement signals are smoothed (filtered) to reduce noise, all state changes and sensor readings are logged over serial, and a button allows manual override when necessary. Additional sensors such as temperature/humidity and flow/level can be integrated, and cloud logging can be extended for remote monitoring and analysis, enabling the end-to-end loop of “measurement → decision → pump control → record” to operate reliably and transparently.

Reference Images

System architecture diagram for smart irrigation — Overall architecture (sensor → MCU → relay/pump → logging)

Actual build photo of the smart farm irrigation prototype — Prototype hardware setup

actual action video

1. Scope

Input (Sensing)

Soil moisture sensor (analog)

Output (Control)

Relay module (Tongling/Songle) → Water pump ON/OFF

Decision Logic

Hysteresis control (e.g., ON ≤ 70%, OFF ≥ 75%) to avoid rapid toggling

2. Hardware Wiring

Module	Pins	Arduino Connection	Note
Soil moisture sensor (analog)	VCC / GND / AOUT	5V / GND / A0	Check water resistance & corrosion; calibrate raw values
Relay module	VCC / GND / IN	5V / GND / D7 (example)	Many modules are LOW-active (invert logic if needed)
Water pump	COM / NO	Drive external PSU (5V/12V) positive lead via relay	Verify current capacity & wire gauge

3. Example Pump Control Code

// Arduino example: Soil moisture → Pump control via relay
int sensorPin = A0;    // Soil moisture sensor (analog)
int relayPin  = 7;     // Tongling relay IN

void setup() {
  pinMode(relayPin, OUTPUT);
  Serial.begin(9600);
  digitalWrite(relayPin, HIGH); // Relay OFF initially (safety)
}

void loop() {
  int raw = analogRead(sensorPin);          // 0~1023 (UNO). ESP32: 0~4095
  int moisture = map(raw, 1023, 0, 0, 100); // Map to percentage (wet → low raw)

  Serial.print("Moisture: ");
  Serial.print(moisture);
  Serial.println(" %");

  // Simple threshold (use hysteresis in production)
  if (moisture <= 70) {
    digitalWrite(relayPin, LOW);   // Relay ON → Pump ON (LOW-active)
  } else {
    digitalWrite(relayPin, HIGH);  // Relay OFF → Pump OFF
  }

  delay(1000); // Check every 1s
}

Note: Some sensors output higher values when wet; adjust map() or threshold logic accordingly. Consider hysteresis (e.g., ON ≤ 70%, OFF ≥ 75%) and minimum on-time for pump protection.

4. Parameter Tuning

Threshold: Start at 70% and fine-tune after field checks
Hysteresis: e.g., ON ≤ 70%, OFF ≥ 75% to prevent frequent switching
Minimum run time: 5–10 s to protect the pump from rapid ON/OFF

5. Safety Checklist

Check relay rating (voltage/current) and wire gauge for the pump load
Add a water level sensor to prevent dry running
Stop immediately on leakage, overheating, or abnormal noise

6. Troubleshooting

Relay works inverted: Confirm LOW-active module; flip HIGH/LOW logic
Noisy readings: Apply averaging or EMA filter; ensure solid ground
Always ON/OFF: Check common ground, separate supplies, and wiring

7. Next Steps (Optional)

Integrate water level / flow sensors
Upgrade to ESP32 for Wi-Fi and remote threshold setting
Log and visualize moisture & pump activity (web dashboard)