IoT Part C: Essay (25%)

Topic: IoT Architecture Applied to Smart Farming

"Design and explain a complete IoT architecture for a Smart Farming system. Include all four layers, justify sensors/actuators, explain Edge/Fog/Cloud, describe protocols, and discuss why this architecture suits precision agriculture." [25 marks]

1. Introduction to Smart Farming (2 marks)

Smart Farming (Precision Agriculture) uses IoT to improve crop production, reduce costs, and minimize environmental impact. Traditional farming uses manual monitoring and fixed schedules, leading to water wastage and unpredictable yields.

An IoT-enabled smart farm uses:

  • Real-time monitoring of soil, weather, and crop health
  • Data-driven decisions for irrigation, fertilization, pest control
  • Automation to reduce labor and error
  • Sustainability through optimized resource usage

Goals: Maximize yield; minimize water/fertilizer waste; enable remote monitoring; predict and prevent diseases.

2. Layer 1: Perception Layer (5 marks)

Interacts with the physical world: collects data with sensors and executes actions with actuators.

Sensors

SensorPurpose
Soil moistureIrrigation decisions; prevent over/under-watering
DHT22 (temp/humidity)Disease risk (e.g. fungal); digital output
LDRSunlight; photosynthesis; ventilation
pH / NPKSoil chemistry; fertilizer type and amount
Rain sensorStop irrigation when raining
CameraAI disease/pest detection

Actuators

ActuatorPurpose
Water pump / solenoidIrrigation; ON when soil moisture < 30%
Ventilation fanGreenhouse temperature; PWM speed
Fertilizer dispenserNutrient delivery based on NPK
LED grow lightsSupplemental light; PWM dimming

Justification: cost-effective, wireless for large farms, long battery life (solar+battery), weather-resistant (e.g. IP65).

3. Layer 2: Transport Layer (4 marks)

Connects devices to the network; transmits data from sensors to gateways and cloud.

TechnologyUse case
LoRaWANLong range (5–15 km), low power, 10-year battery; rural areas
Wi-FiGreenhouse; cameras; high bandwidth
Zigbee / BLE meshShort range, low power, mesh
4G/5GBackup; critical alerts

MQTT (primary): Lightweight, pub/sub, QoS, works over LoRa/WiFi/cellular. Topics e.g. Farm/Greenhouse1/Sensor/SoilMoisture; dashboard subscribes to Farm/+/Sensor/#.

[Sensors] --LoRa--> [Gateway] --WiFi/4G--> [Edge]
                                      |--MQTT--> [Cloud]
                                      |--Local--> [Fog]

4. Layer 3: Processing Layer (5 marks)

Edge, Fog, and Cloud divide the work.

  • Edge (on device, e.g. ESP32): Filter data (e.g. 5-min average); local rules (if moisture < 30% → pump on); anomaly detection. Saves bandwidth, fast response, works offline.
  • Fog (gateway / on-site server): Aggregate sensors; rules engine (e.g. humidity > 85% + temp > 30°C → alert); OTA updates; lightweight AI (e.g. disease from camera). Low latency, privacy, works if internet fails.
  • Cloud: Long-term storage (time-series, SQL); big data analytics; ML training (yield, disease models); multi-farm analytics. Scalable storage and compute.
TaskEdgeFogCloud
Emergency pump off
Daily irrigation schedule
Yield prediction / train AI

5. Layer 4: Application Layer (4 marks)

Dashboard (web + mobile): real-time graphs, map, camera; manual pump override; irrigation schedule; alerts (critical → SMS+push+email; warning → app). AI: predictive maintenance, yield forecast, recommendations (planting date, fertilizer reduction).

6. Why This Architecture Fits Smart Farming (3 marks)

  • Scalability: 10 → 1,000 sensors; cloud scales with data.
  • Cost: Edge reduces data transfer; low-power sensors; open-source (MQTT, InfluxDB).
  • Reliability: Edge works offline; fog backup; mesh self-healing.
  • Real-time: Edge activates pump in < 5 s; cloud-only would be 30–60 s.
  • Energy: LoRa 10-year battery; solar gateway; pump only when needed.
  • Sustainability: Precision irrigation and fertilization reduce waste and runoff.

7. Conclusion (2 marks)

The 4-layer architecture integrates Perception (sensors/actuators), Transport (LoRa/WiFi/MQTT), Processing (Edge/Fog/Cloud), and Application (dashboards, AI). It addresses large spread (LoRa), limited power (edge + solar), immediate action (local control), and long-term planning (cloud). Result: scalable, cost-effective, sustainable solution that increases yield and reduces resource use.

Essay structure: Intro 2 · Perception 5 · Transport 4 · Processing 5 · Application 4 · Justification 3 · Conclusion 2 = 25 marks.