Computer Engineering is coming soon

We're still writing and reviewing the notes for every subject, and we'll open it when every subject is ready. Meanwhile you can practise aptitude, ask Aria about any subject, or see the branches that are ready.

IoT applications: environment, agriculture, industry and health

Practical notes on IoT applications in environment, agriculture, industry and health: requirements, typical system designs, deployment math (coverage, battery life), examples and exam guidance.

Drafted with Aria, reviewed by the AiCanCode.org team · Updated . How we write and check our notes. Spotted an error? Use Give Feedback at the bottom of the page.

Why it matters

IoT applications let engineers collect sensor data and control actuators automatically in real-world domains. An engineer uses these designs to choose sensors, design nodes and gateways, size batteries, plan deployments (how many nodes and where), and meet constraints such as latency, reliability and power for environment, agriculture, industry and health projects.

Key ideas

  • Actuator: a device that changes the physical environment (valve, motor). Defined as the complement of a sensor; it receives commands from MCU and changes state.
  • Gateway: device that bridges low-power local networks to the internet/cloud. Usually does protocol translation and aggregates data.
  • Duty cycle: fraction of time a node is in active mode (transmitting, sensing, computing) versus sleeping. Duty cycle = active time / total time; used to estimate average power. In unlicensed bands, regulators may also impose a transmit duty-cycle limit (e.g., ~1% in EU 868 MHz), which constrains how often a node may transmit.
  • Sampling rate: how often a sensor takes a measurement, in samples per second (Hz) or samples per minute/hour. Higher sampling gives better temporal resolution but increases energy and storage use.
  • Coverage: area or volume where a sensor effectively measures the physical variable. For ground sensors, often approximated by a radius r (m) around the node.
  • Latency: time between event occurrence and its detection/notification. Important in health (alarm) and industry (control).
  • Reliability: probability the system correctly delivers data within required latency. Often expressed as packet delivery ratio or uptime.

Key constraints and how they differ by domain (short):

  • Environment: wide areas, low-power, long lifetime, tolerance for higher latency, sparse deployments; often energy harvesting.
  • Agriculture: battery-powered or energy-harvested nodes; moderate latency; robustness to dust/water; coverage planning for fields.
  • Industry: strict latency, high reliability, often wired power available; real-time control; often uses local gateways and redundant paths.
  • Health: strong security and privacy requirements; low-latency alarm delivery; redundancy for safety; regulatory compliance.

Design steps (first principles):

  1. Specify functional requirements: variables measured, sampling rate, actuation need, latency, reliability, security.
  2. Choose sensors and actuators that meet accuracy and environmental rating (e.g., IP rating for water/dust).
  3. Size communication: data rate, packet frequency, protocol (refer to topics: MQTT, CoAP, LoRaWAN, 802.15.4, BLE) depending on range and duty cycle.
  4. Size power: compute average current from duty cycle, choose battery or energy harvesting.
  5. Plan deployment: compute number of nodes for coverage, sensor placement strategy (grid/hex/cluster), and gateway locations.
  6. Design software: edge filtering, event detection, secure communication to cloud.

Where to draw diagrams in an exam:

  • Sensor node block diagram (see above). Label: sensors, ADC (if present), MCU, memory, radio, power, optional actuator.
  • Deployment map: draw rectangular field, mark sensor positions in hexagonal grid, show gateway position and radius of reliable link.

Assumptions to state in exam answers: environmental conditions, battery nominal voltage, ideal sphere/circle coverage (if used), uniform field and isotropic sensing radius, negligible interference unless specified.

Formulas

Duty cycle D = t_active / T_cycle

  • D: duty cycle (dimensionless), t_active: active time per cycle (s), T_cycle: cycle period (s). Applies when node repeats identical sleep/active cycles.

I_avg = I_active·(t_active/T_cycle) + I_sleep·(1 - t_active/T_cycle)

  • I_avg: average current (A), I_active: current in active mode (A), I_sleep: current in sleep (A). Use when alternating active and sleep states.

Battery life (hours) = (Q_bat (Ah)) / I_avg (A)

  • Q_bat: battery capacity in ampere-hour (Ah). Gives estimated lifetime ignoring self-discharge and temperature effects.

E_tx = V_bat · I_tx · t_tx

  • E_tx: energy used for one transmission (J), V_bat: battery voltage (V), I_tx: current during transmission (A), t_tx: transmission duration (s). Use for per-packet energy calculation.

Area_per_sensor_hex = (3·√3 / 2) · r^2

  • r: coverage radius (m), area in m^2. Use for hexagonal covering (tight coverage without gaps) of a 2D plane; here r is the distance from the node to a hexagon vertex (circumradius).

N_sensors = ceil(Area_total / Area_per_sensor_hex)

  • Number of sensors required (integer).

Duty cycle (%) = D·100 — use to report percentage.

Notes: always carry SI units. These formulas assume ideal behaviours (constant currents, no temperature dependency, isotropic sensing and radio range).

Worked examples

Example 1 — Battery life for an agricultural soil-moisture node Problem: A soil-moisture sensor node uses a ~3.0 V nominal battery pack of 2 × AA alkaline cells in series, rated 2.4 Ah. Each measurement cycle (every 30 minutes) the node wakes, measures (0.5 s at I_active = 15 mA), processes (0.5 s at 12 mA), and transmits a packet (2 s at I_tx = 50 mA). Sleep current I_sleep = 50 µA. Estimate battery life in months. Steps:

  1. Cycle period T_cycle = 30 minutes = 1800 s. Active time t_active = 0.5 + 0.5 + 2 = 3.0 s.
  2. Compute average current: I_avg = I_active_equiv·(t_active/T_cycle) + I_sleep·(1 - t_active/T_cycle). Here active currents vary during active phase; compute energy-weighted average during active phase: I_active_equiv = (0.5·15 + 0.5·12 + 2·50) / 3.0 = (7.5 + 6 + 100) / 3 = 113.5 / 3 = 37.833... mA.
  3. Now I_avg = 0.037833 A·(3/1800) + 0.00005 A·(1 - 3/1800) = 0.037833·0.0016667 + 0.00005·0.9983333 ≈ 0.00006306 + 0.00004992 = 0.00011298 A ≈ 0.11298 mA.
  4. Battery capacity Q_bat = 2.4 Ah. Battery life (hours) = 2.4 / 0.00011298 ≈ 21244 hours. Convert to months: 21244 h / (24·30) ≈ 29.5 months ≈ 29.5 months. Answer: Battery life ≈ 29.5 months (ideal conditions). Note: actual lifetime will be less due to self-discharge, temperature and battery discharge curves.

Example 2 — Number of air-quality sensors for a 1 km^2 campus Problem: Cover a rectangular campus of 1.0 km^2 (1,000,000 m^2) using sensors with effective sensing radius r = 50 m, using hexagonal packing. Steps:

  1. Area per sensor (hex) = (3·√3 / 2)·r^2. Compute √3 ≈ 1.732, so (3·1.732/2) ≈ 2.598. Area_per ≈ 2.598·50^2 = 2.598·2500 = 6,495 m^2.
  2. N_sensors = ceil(1,000,000 / 6,495) = ceil(153.96) = 154 sensors. Answer: 154 sensors (approx.). State assumptions: isotropic sensing, obstacles ignored, sensors placed optimally in hex grid.

Example 3 — Model long-answer: Compare IoT requirements across environment, agriculture, industry, health (exam-style answer) Write a structured comparison: define each domain, list three main requirements (power, latency, reliability), then explain typical hardware and network choices, and give one short example deployment and its special considerations (e.g., environmental sensors may use LoRaWAN and solar; industrial may use wired power plus redundant Ethernet/Time-Sensitive Networking (TSN)). Conclude with security/privacy differences (health highest). A full-marks answer states explicit numeric constraints when possible (e.g., health alarm latency < seconds, industry control latency < tens of ms), and justifies choices.

Common mistakes

  • Assuming circular coverage without accounting for packing inefficiency; using πr^2 underestimates required sensors for non-overlapping optimal packing.
  • Ignoring sleep current; for very low duty cycles it dominates lifetime calculation.
  • Mixing current (A) and energy (J) formulas without voltage; remember E = V·Q or E = V·I·t.
  • Not stating assumptions (temperature, battery ageing, obstacles) in exam answers.

For your exam

Typical question types: definitions (define duty cycle, gateway, actuator), short notes (compare requirements across domains), derivations (derive battery life formula from duty cycle), numericals (battery life, number of sensors for coverage), diagrams (sensor node block diagram, deployment map), design problems (choose sensors and network for a use-case and justify). What full-marks answers include:

  • Clear definitions with bold terms.
  • Stated assumptions and units.
  • Step-by-step derivation or calculation, unit-carrying arithmetic, and a final clearly highlighted result.
  • Diagrams labeled correctly when asked.
  • Discussion of trade-offs (energy vs latency, cost vs coverage, reliability vs power).

Quick check

  1. Define duty cycle in one sentence.
  2. Using hexagonal packing, what is the area served by one sensor with sensing radius 40 m? (give numeric result).
  3. A node wakes for 2 s every hour; active current 30 mA, sleep current 20 µA, battery 1.2 Ah. Estimate battery life in days. Answers: Duty cycle = active time divided by total cycle period. Area = (3·√3/2)·40^2 ≈ 2.598·1600 ≈ 4,156.8 m^2. Battery life: T_cycle = 3600 s, t_active = 2 s, I_avg ≈ 0.03·(2/3600) + 0.00002·(1 - 2/3600) ≈ 0.00001667 + 0.00001999 ≈ 0.00003666 A → life = 1.2 / 0.00003666 ≈ 32738 h ≈ 1364 days ≈ 3.74 years.

(End of notes.)

Finished this topic? Mark it so your progress, study plan and readiness keep up.

Stuck on something here?