59  Processes & Systems: Fundamentals

In 60 Seconds

IoT systems are modeled as processes that transform inputs (sensor readings) into outputs (actuator commands) through defined algorithms. Understanding block diagrams, feedback loops, and control types (open-loop vs closed-loop) is essential for designing stable, self-regulating IoT systems.

59.1 Learning Objectives

By the end of this series, you will be able to:

• Distinguish between processes (input-output transformations) and systems (interconnected components) in IoT architectures
• Differentiate open-loop and closed-loop control systems and identify when each is appropriate
• Trace feedback loop signal flow from sensor measurement through error calculation to actuator response
• Explain how Proportional, Integral, and Derivative terms contribute to precise control
• Evaluate trade-offs between control strategies (cost, complexity, accuracy) for specific IoT scenarios
• Design self-regulating systems that automatically adjust output to maintain a desired setpoint

59.2 Prerequisites

Before diving into this series, you should be familiar with:

• Sensor Fundamentals and Types: Understanding how sensors measure physical phenomena is essential for grasping how systems transform inputs into meaningful data
• Actuators: Knowledge of actuators helps understand how systems produce physical outputs in response to control signals
• IoT Reference Models: Familiarity with layered IoT architectures provides context for where processes operate within the device-to-cloud stack

Key Concepts

• Process: A system with inputs, internal states, and outputs that undergoes transformation over time — in IoT, this includes physical processes (temperature heating) and computational processes (data filtering).
• Open-Loop Control: A control system that applies a fixed input based on a model without measuring the actual output — fast and simple, but cannot correct for disturbances or model errors (e.g., a fixed-duration sprinkler timer).
• Closed-Loop Control: A control system that measures the actual output and feeds it back to adjust the input — can compensate for disturbances and achieve precise set-point tracking (e.g., a thermostat).
• Set Point: The desired target value for a controlled variable — the goal the controller is trying to achieve (e.g., 22°C room temperature, 60% soil moisture).
• Error Signal: The difference between the set point and the measured process variable — the fundamental input to all feedback controllers: e(t) = r(t) - y(t).
• Disturbance: An external input that affects the process but is not controlled — in IoT systems, disturbances include weather changes, user behavior, equipment wear, and sensor drift.
• Transfer Function: A mathematical model relating the output of a process to its input — used to predict system behavior and design controllers without building physical prototypes.

59.3 Minimum Viable Understanding

• Every IoT device is a system: It takes inputs (sensor readings, user commands), applies a process (algorithms, logic, control), and produces outputs (actuator actions, data transmissions) – understanding this input-process-output framework is the single most important mental model for IoT design.
• Closed-loop control is what makes IoT “smart”: The difference between a dumb timer and an intelligent system is feedback – measuring the actual output and automatically adjusting inputs to reach a desired setpoint.
• PID is the workhorse control algorithm: Proportional-Integral-Derivative control handles the vast majority of real-world IoT regulation tasks (temperature, speed, position), and understanding its three components lets you tune any feedback system.

Sensor Squad: How Processes Work – A Pizza Kitchen Adventure!

Sammy the Sensor gathered the squad around a table. “Today we’re learning about processes and systems – and we’re going to use a pizza kitchen to explain it!”

Lila the Light Sensor said: “A system is like the entire pizza kitchen – the oven, the ingredients, the chef, and the recipe all working together. A process is what the kitchen does: it takes raw dough and toppings (inputs) and turns them into a delicious pizza (output)!”

Max the Motion Detector added: “But here’s the cool part – what if the pizza comes out burnt? In an open-loop kitchen, the chef just sets the oven to 200 degrees and walks away. Hope for the best! But in a closed-loop kitchen, the chef keeps checking the pizza and turns the heat up or down. That’s called feedback – and it’s what makes IoT devices smart!”

Bella the Buzzer finished: “And PID control is like having a super-chef! The P part says ‘the pizza is too cold, turn up the heat NOW.’ The I part says ‘the pizza has been too cold for a while, turn it up MORE.’ And the D part says ‘whoa, the temperature is rising too fast, ease off a little!’ Together they bake the perfect pizza every time!”

Key Ideas for Young Engineers:

Pizza Kitchen	IoT System	Example
Ingredients going in	Inputs (sensor data)	Temperature reading
Cooking the pizza	Process (algorithm)	Compare to setpoint
Finished pizza	Output (actuator action)	Turn heater on/off
Chef checking the pizza	Feedback (sensor loop)	Thermocouple in oven
Adjusting the oven dial	Control (PID)	Smart thermostat

59.4 For Beginners: What Are Processes and Systems?

Simple Explanation

Analogy: Think of a vending machine. You put in money and press a button (inputs), the machine selects and dispenses your snack (process), and you get a drink or candy bar (output). The entire vending machine is the system; the steps it follows to deliver your snack are the process.

Three things to remember:

1. A system is a collection of parts working together toward a goal (the vending machine hardware, software, and mechanisms)
2. A process is the transformation the system performs – turning inputs into outputs (coins + button press into a dispensed snack)
3. Control is how the system decides what to do – either following fixed rules (open-loop) or adjusting based on what actually happens (closed-loop)

Why does this matter for IoT? Every smart device – from a fitness tracker to a factory robot – follows this pattern. Once you see the input-process-output structure, you can design, debug, and improve any IoT system.

59.5 Chapter Overview

About This Series

This chapter introduces the fundamental concepts of processes and systems in IoT architectures. Understanding how electronic systems transform inputs into outputs through controlled processes is essential for designing robust IoT solutions.

This topic has been organized into three focused chapters:

1. Core Definitions - What are processes and systems, block diagrams, decomposition
2. Control Types - Open-loop vs closed-loop, feedback mechanisms
3. PID Control - Proportional-Integral-Derivative control theory and applications

59.6 The Big Picture: How Processes Fit Into IoT Architecture

Before diving into the individual chapters, it is important to understand how processes and control systems relate to the broader IoT architecture. The following diagram shows where process control sits within a typical IoT system.

59.7 Chapter 1: Core Definitions

Processes & Systems: Core Definitions

Learn the foundational concepts that underpin all IoT system design:

• System vs Process: Understand the difference between the collection of components (system) and the transformation they perform (process)
• Block Diagram Representation: Use abstraction to represent complex systems as interconnected black boxes
• IoT System Decomposition: Break down systems into hardware, software, and network subsystems
• Input-Output Transformations: Analyze how sensors provide inputs, microcontrollers process data, and actuators produce outputs

Key Concept: Every IoT device can be analyzed as a system with inputs, a process, and outputs – understanding this framework is essential for debugging and optimization.

The following diagram illustrates the core input-process-output model that applies to every IoT device:

59.8 Chapter 2: Control Types

Processes & Systems: Control Types

Explore the fundamental distinction between control strategies:

• Open-Loop Control: Systems that operate based on predetermined inputs without measuring output
• Closed-Loop Control: Systems that continuously measure output and adjust to maintain desired setpoint
• Feedback Mechanisms: How sensors, controllers, and actuators work together in feedback loops
• Real-World Applications: Smart factory temperature control, irrigation systems, HVAC automation

Key Concept: Closed-loop systems that measure output and adjust inputs automatically are the foundation of “smart” IoT – turning simple timers into intelligent devices.

The diagram below contrasts open-loop and closed-loop control, showing how feedback transforms a simple system into a smart one:

59.9 Chapter 3: PID Control

Processes & Systems: PID Control

Master the most widely used feedback control algorithm:

• The PID Formula: Proportional, Integral, and Derivative control actions
• Tuning Controllers: Finding optimal gain values for system performance
• Real IoT Examples: Smart thermostats, drone stability, water tank level control
• Control Strategy Selection: When to use PID vs simpler on/off control

Key Concept: PID combines three actions – react to current error (P), eliminate persistent offset (I), and dampen oscillations (D) – to achieve precise, stable control.

This diagram shows how the three PID components work together in a feedback loop:

59.10 Control Strategy Decision Flowchart

Not sure which control strategy to use for your IoT project? Follow this decision flowchart:

59.11 Quick Reference: Control System Comparison

Control Type	Feedback	Accuracy	Complexity	IoT Example
Open-Loop	None	Low	Simple	Timer-based irrigation
Closed-Loop (On/Off)	Yes	Medium	Moderate	Simple thermostat
Closed-Loop (PI)	Yes	High	Moderate-High	Motor speed control
Closed-Loop (PID)	Yes	Very High	Higher	Smart HVAC, drones

Quick Calculation: On/Off vs PID Energy Efficiency

viewof room_area_m2 = Inputs.range([20, 500], {value: 200, step: 10, label: "Greenhouse area (m²)"})

viewof heater_kw = Inputs.range([1, 50], {value: 10, step: 1, label: "Heater power (kW)"})

viewof hours_per_day = Inputs.range([4, 24], {value: 12, step: 1, label: "Operating hours/day"})

{
  const kwh_per_day_on_off = heater_kw * hours_per_day * 0.65; // ~65% duty cycle for on/off
  const kwh_per_day_pi = heater_kw * hours_per_day * 0.50;     // ~50% duty cycle for PI
  const savings_kwh = kwh_per_day_on_off - kwh_per_day_pi;
  const savings_pct = ((savings_kwh / kwh_per_day_on_off) * 100).toFixed(1);
  const annual_savings_kwh = savings_kwh * 365;
  const annual_savings_usd = (annual_savings_kwh * 0.12).toFixed(0);
  return html`<div style="background: var(--bs-light, #f8f9fa); padding: 1rem; border-radius: 8px; border-left: 4px solid #16A085; margin-top: 0.5rem;">
    <p style="margin:0"><strong>On/Off control:</strong> ~${kwh_per_day_on_off.toFixed(1)} kWh/day (65% duty cycle typical)</p>
    <p style="margin:0.3rem 0 0"><strong>PI control:</strong> ~${kwh_per_day_pi.toFixed(1)} kWh/day (50% duty cycle typical)</p>
    <p style="margin:0.3rem 0 0; border-top: 1px solid #dee2e6; padding-top: 0.3rem"><strong>Annual energy savings:</strong> ${annual_savings_kwh.toFixed(0)} kWh (~${savings_pct}%, ~$${annual_savings_usd}/year at $0.12/kWh)</p>
    <p style="margin:0.3rem 0 0; font-size:0.85em; color:#7F8C8D;">Note: Actual savings depend on climate and setpoint. PI control reduces overshoot, cutting wasted heating cycles.</p>
  </div>`
}

59.12 Worked Example: Designing a Smart Greenhouse Climate System

Scenario: A small farm wants to automate its 200 m² greenhouse. The crops require temperature between 20–25 degrees C and humidity between 60–80%. The farm has a limited budget and the staff has no control engineering background.

59.12.1 Step 1: Identify the System

Component	Details
Inputs	Temperature sensor (DHT22), humidity sensor (DHT22), light sensor (BH1750), user setpoint via app
Process	ESP32 microcontroller running control algorithm
Outputs	Heating mat relay, exhaust fan relay, misting nozzle relay, grow light relay
Disturbances	Outdoor temperature changes, door opening, cloud cover

59.12.2 Step 2: Choose Control Strategy Per Output

Temperature control (heating mat + exhaust fan):

• The output must match a setpoint (20–25 degrees C) – closed-loop needed
• The DHT22 can measure temperature – feedback available
• Temperature oscillation of +/- 2 degrees C is acceptable – on/off control with hysteresis is sufficient
• Decision: On/off control with 2 degree C hysteresis band (heater ON below 20 degrees C, OFF above 22 degrees C; fan ON above 25 degrees C, OFF below 23 degrees C)

Putting Numbers to It

Tracking greenhouse temperature with on/off control (2°C hysteresis):

Morning cold start (8:00 AM, outdoor=5°C):

• Temp=15°C → Heater ON (below 20°C threshold)
• 9:00 AM: Temp=20°C → Heater still ON (hasn’t reached 22°C upper limit)
• 9:30 AM: Temp=22°C → Heater OFF (crossed upper threshold)
• Passive cooling to 20.5°C over 45 min → Heater ON again (below 20°C)
• Oscillates 20-22°C with ~60 min cycle

Afternoon sun heating (2:00 PM, outdoor=28°C):

• Temp=23°C → Fan OFF (below 25°C threshold)
• 2:30 PM: Temp=25°C → Fan ON (crossed lower cooling limit)
• 3:00 PM: Temp=23°C → Fan still ON (hasn’t dropped to 23°C cutoff)
• 3:30 PM: Temp=23°C → Fan OFF (reached lower threshold)
• Oscillates 23-25°C with ~90 min cycle

Wider hysteresis (4°C) → fewer relay cycles (longer component life), but bigger swings (18-22°C, 23-27°C). Tighter (1°C) → smoother temp, but 2x relay wear. 2°C is the sweet spot for crops.

Humidity control (misting nozzle):

• Must maintain 60–80% range – closed-loop needed
• DHT22 can measure humidity – feedback available
• Misting is either on or off (no variable rate) – on/off control is the only option
• Decision: On/off control (mist ON below 60%, OFF above 75%)

Light control (grow lights):

• Crops need 14 hours of light per day – schedule-based
• Could add light sensor for cloudy day compensation, but budget is limited
• Decision: Open-loop timer (lights ON 6:00 AM to 8:00 PM) with optional light sensor upgrade later

59.12.3 Step 3: Result

The system uses a mix of control strategies matched to each output’s requirements and the available hardware. No PID control was needed because the acceptable tolerance bands are wide enough for on/off control. The total cost stays within budget by avoiding expensive variable-rate actuators.

Key Lesson: Not every IoT system needs PID. Choose the simplest control strategy that meets your accuracy requirements – over-engineering adds cost, complexity, and potential failure points.

59.13 Common Pitfalls

Common Pitfalls in IoT Process & Control Design

1. Using open-loop when feedback is available: Many beginners deploy timer-based irrigation even when soil moisture sensors cost only a few dollars. If you can measure the output, always prefer closed-loop control – the reliability improvement is almost always worth the added sensor cost.

2. Jumping straight to PID control: PID is powerful but adds complexity in tuning, potential for oscillation, and debugging difficulty. For many IoT applications (thermostats, irrigation, lighting), simple on/off control with hysteresis is more robust and easier to maintain. Always ask: “Is my required precision tight enough to justify PID?”

3. Ignoring disturbances in system design: A control system designed for a sealed lab may fail in a greenhouse with opening doors or a factory with varying loads. Always identify external disturbances during the design phase and ensure your control strategy can handle them.

4. Forgetting actuator limitations: PID assumes the actuator can produce any output level (proportional control). If your actuator is binary (relay-based heater: ON or OFF), PID offers no advantage over on/off control with hysteresis. Match your control algorithm to your actuator capabilities.

5. Setting the sampling rate too low: If the control loop reads the sensor once per minute but the process can change significantly in 10 seconds (e.g., motor speed), the controller will always be reacting to stale data. The sampling rate must be at least 5–10 times faster than the process time constant.

59.14 Knowledge Check

Test your understanding of the key concepts covered in this series overview.

## Real-World Case Study: Nest Learning Thermostat – From On/Off to Predictive Control {#arch-process-nest-case-study}

Case Study: Google Nest’s Control Evolution (2011–2024)

The Nest Learning Thermostat illustrates how control strategy selection drives product success in consumer IoT.

Generation 1 (2011): Adaptive On/Off Control

Nest started with hysteresis-based on/off control (heater ON below setpoint - 1 degree C, OFF above setpoint + 1 degree C) but added a learning layer. The thermostat recorded when users adjusted temperature and after 1–2 weeks built a daily schedule automatically. This was still fundamentally on/off control with a learned schedule – no PID.

Metric	Before Nest (Manual Thermostat)	Nest Gen 1
Average daily adjustments by user	4.2	0.3 (after learning)
Energy savings vs fixed schedule	Baseline	10–12%
Overshoot (heating)	+/- 2 degree C	+/- 1.5 degree C
User satisfaction (J.D. Power)	62/100	81/100

Generation 3+ (2015–2024): Predictive Pre-Conditioning

Later generations added predictive control that anticipated heating/cooling needs based on outdoor temperature trends, home thermal mass, and HVAC system response time. The system learned that a particular home takes 45 minutes to heat from 18 to 22 degrees C, so it starts heating at 6:15 AM instead of 7:00 AM.

This is effectively a simplified PI controller: - Proportional: Larger temperature gap = earlier start time - Integral: Accumulates “the house is slow to heat” learning over weeks, permanently adjusting lead times

Results from Nest’s published energy studies (2015):

• Average 10–12% savings on heating, 15% on cooling across 735 homes
• Homes in mild climates (California): 5–8% savings
• Homes in extreme climates (Minnesota): 12–18% savings
• Key insight: The learning algorithm matters more than the control algorithm. Users who overrode the schedule frequently saw only 3% savings, while hands-off users saw 20%+.

Design Lesson: Nest succeeded not because of sophisticated PID control (it uses on/off with prediction), but because it matched control complexity to the actuator (binary relay furnace) while adding value through schedule learning. This validates the chapter’s core principle: choose the simplest control strategy that meets your accuracy requirements.

Match Control Strategies to Application Characteristics

Order the IoT Control System Design Workflow

Label the Diagram

💻 Code Challenge

59.15 Summary

This chapter series covers the foundational concepts that underpin all intelligent IoT system design.

Core Concepts Covered:

Concept	Key Insight	Chapter
System	A collection of components working together toward a goal	Core Definitions
Process	The transformation that converts inputs to outputs	Core Definitions
Open-Loop	Fixed rules, no feedback – simple but inaccurate	Control Types
Closed-Loop	Measures output and adjusts – the basis of “smart”	Control Types
On/Off Control	Simplest feedback – sufficient for wide tolerances	Control Types
PID Control	Three-component algorithm for precise regulation	PID Control

Key Takeaways:

1. Every IoT device can be analyzed as inputs + process + outputs. This mental model is your most powerful debugging tool.
2. Feedback is the dividing line between a simple device and a smart one. If you can measure the output, close the loop.
3. Match control complexity to requirements: Open-loop for schedules, on/off for wide tolerances, PI for most precision tasks, full PID only when overshoot must be minimized.
4. Match control to actuator capabilities: PID with a binary relay is wasted complexity. Proportional control requires a proportional actuator.
5. Disturbances are unavoidable in real-world IoT deployments. Always design your control strategy to handle them, not just ideal conditions.

59.16 What’s Next

If you want to…	Read this
Study core system definitions	Core Process System Definitions
Learn about control types	Control Types
Explore PID control fundamentals	PID Control Fundamentals
Study the process decision guidance	Process Decision Guidance
Practice with labs and review	Labs and Review