Explain why energy is the fundamental limiting resource in IoT design
Define power and energy in the context of IoT devices
Identify the key tradeoffs in energy-aware IoT system design
Explain the gap between battery technology improvement and computing demands
Apply basic power and energy concepts to estimate device lifetime
In 60 Seconds
Energy is the most constrained resource in IoT design because batteries improve only 5–8% per year while computing demands grow 40% per year — making energy-aware design the critical engineering skill that separates IoT products lasting weeks from those lasting years.
Key Concepts
Energy Budget: Total energy available to an IoT device over its operational lifetime, typically measured in mAh (battery capacity × voltage); must cover all sensing, processing, and communication operations.
Power Consumption: Rate of energy use measured in watts or milliwatts; varies dramatically between active (10-100 mA) and sleep (1-10 µA) states in IoT devices.
Sleep Mode: Low-power device state used between sensing events; reducing sleep current from 100 µA to 1 µA can extend battery life 100x at low duty cycles.
Duty Cycle: Fraction of time a device spends in active (power-consuming) mode; reducing duty cycle from 10% to 1% approximately 10x extends battery life.
Battery Chemistry: Electrochemical characteristics of batteries affecting energy density, discharge profile, temperature performance, and self-discharge rate; must match deployment requirements.
Energy Harvesting: Collecting ambient energy (solar, vibration, RF) to supplement or replace batteries; requires careful power management to match harvested power to consumption.
Coulomb Counting: Battery state-of-charge estimation technique measuring charge in and out of battery; used for accurate battery life prediction in IoT applications.
2.3 For Beginners: Why Energy Matters in IoT
Imagine you have a tiny sensor that monitors the temperature in your garden. It needs to:
Wake up every hour
Read the temperature
Send the data to your phone
Go back to sleep
Each of these steps uses energy from a small battery. If the battery dies, your sensor stops working. That’s why every IoT designer obsesses over every microamp of current - it’s the difference between a device that lasts weeks versus years!
Key concepts:
Power = how fast you use energy (like water flow from a tap)
Energy = total amount used over time (like total water in a bucket)
Battery Life = Energy stored / Average power consumption
For Kids: Meet the Sensor Squad!
Sammy the Sensor says: “I’m like a night owl - I only wake up when something interesting happens! The rest of the time, I’m snoozing in deep sleep mode, using almost no battery. That’s how I can stay awake for YEARS on a tiny coin battery!”
Lila the LED explains: “I use 20 milliamps when I’m on - that’s like drinking through a straw. But if I stay on all day, I’ll drain the battery super fast! That’s why smart devices only light me up for a second.”
Max the Microcontroller shares: “I have different ‘sleep modes’ like you have different levels of napping! Light sleep is like dozing on the couch. Deep sleep is like being in bed with the covers over my head - I use almost zero energy but take longer to wake up!”
Think about it like this: Imagine your phone’s battery is a juice box. If you play games ALL day, the juice runs out fast. But if you only check your phone sometimes and let it sleep the rest of the time, the juice lasts much longer! That’s exactly how IoT devices work - they sleep most of the time to save their battery “juice”!
This tutorial demonstrates practical deep sleep implementation on ESP32, showing how to achieve under 10 microamps sleep current for extended battery operation.
Video: Battery Optimization Strategies for IoT Devices
Coverage:
Battery selection criteria for IoT applications
Duty cycling fundamentals and calculation
Low-power communication protocols comparison
Energy harvesting integration basics
Estimated Watch Time: 20 minutes
2.6 Introduction
Energy is the fundamental limiting resource in IoT design. While processing power, memory, and communication bandwidth have improved exponentially following Moore’s Law, battery technology has progressed at a much slower pace—approximately 5-8% per year. This disparity creates the central challenge of IoT system design: how to maximize device functionality while operating within strict energy constraints.
The consequences of poor energy management are significant. A device that requires frequent battery replacement creates maintenance costs that may exceed the device’s value. For deployments in remote locations—agricultural fields, wildlife habitats, industrial facilities—battery replacement may be impractical or impossible. The dream of “deploy and forget” IoT sensors lasting 5-10 years requires exceptional energy discipline.
2.7 Definition
Energy-aware IoT design encompasses the systematic consideration of energy consumption at every level of the system—from component selection through circuit design, firmware implementation, and communication protocols. The goal is to maximize useful work performed per unit of energy consumed, extending battery life while meeting application requirements.
Power represents the instantaneous rate of energy consumption, measured in watts (W):
\[P = V \times I\]
Where V is voltage and I is current.
Energy represents total consumption over time, measured in joules (J) or watt-hours (Wh):
\[E = P \times t = V \times I \times t\]
For battery-powered devices, we typically work with milliamp-hours (mAh), where battery life is:
\[\text{Battery Life (hours)} = \frac{\text{Battery Capacity (mAh)}}{\text{Average Current (mA)}}\]
2.8 Tradeoff: Primary Battery vs Rechargeable + Harvesting
Factor
Primary Battery
Rechargeable + Harvesting
Initial Cost
Lower
Higher (harvester + charger + battery)
Maintenance
Periodic replacement
None (if sized correctly)
Reliability
Predictable lifetime
Weather/environment dependent
Complexity
Simple
Complex (MPPT, BMS, sizing)
Best For
Indoor, accessible locations
Outdoor, remote, permanent installations
2.9 Tradeoff: Edge Processing vs Cloud Offloading
Factor
Edge Processing
Cloud Offloading
Radio Energy
Minimal (send results only)
High (send raw data)
Processing Energy
Higher (local compute)
Minimal
Latency
Low (local decision)
High (network round-trip)
Break-even
100-1000 operations = 1 byte TX
Simple filter/aggregate locally
Best For
Complex analytics, privacy-sensitive
Simple sensors, cloud-dependent apps
2.10 Tradeoff: Aggressive Sleep vs Always-On Responsiveness
Factor
Aggressive Sleep (99%+)
Always-On
Battery Life
Years
Days to weeks
Response Latency
100ms - seconds
Immediate
Wake-up Energy
Overhead per wake
None
Complexity
State management, context restore
Simple
Best For
Periodic monitoring, thresholds
Real-time control, continuous stream
Energy Harvesting is Not a Free Lunch
Energy harvesting (solar, thermal, RF, kinetic) sounds appealing—“free” energy from the environment! But there are critical pitfalls:
Indoor solar is usually NOT viable: Office lighting provides only 5-10 µW/cm². A typical IoT device needs 10-100× more just for sleep current. Indoor solar only works in very bright environments (greenhouses, near windows).
Outdoor solar requires careful sizing: You need enough panel area AND battery capacity to handle:
Cloudy days (3-7 days in many climates)
Winter reduced sunlight (50% of summer in temperate zones)
Panel degradation over time
Energy storage is still required: Even with energy harvesting, you need a battery or supercapacitor to buffer between harvesting peaks and consumption needs.
Thermal/kinetic/RF harvesting is niche: These sources produce microwatts. They work for specific applications (body heat, vibration) but rarely as primary power sources.
Rule of Thumb: Design your IoT device to work on battery alone first. Add energy harvesting as a battery life extender, not a replacement for good power management.
Figure 2.1: Energy flow in a typical IoT device showing power sources, management, and consumers
2.11 The Fundamental IoT Constraint: Battery Gap is Widening, Not Closing
Figure 2.2: The growing gap between computing advancement and battery technology improvement
The chart above illustrates the fundamental challenge: while processors have become 10,000× more powerful since 1990, batteries have only improved by approximately 10×. This means:
Software cannot assume abundant energy - Even as devices get more capable, energy efficiency must remain paramount
Sleep modes are essential - Devices must spend most of their time in ultra-low-power states
Communication is the dominant cost - Wireless transmission consumes 100-1000× more energy than local processing
Every milliamp matters - Design decisions that save 100 uA can extend battery life by months
Common Pitfalls in Energy-Aware Design
Pitfall 1: Ignoring Sleep Current
Many designers focus only on active mode current. But if your device sleeps 99% of the time, even 100uA sleep current dominates the power budget!
ESP32 light sleep: 800uA (bad for battery life)
ESP32 deep sleep: 10uA (acceptable)
ESP32 hibernation: 5uA (best for long deployments)
Pitfall 2: Forgetting Peripheral Leakage
Sensors, LEDs, and pull-up resistors consume power even when “off”:
LED with 10k resistor at 3.3V: 330uA constantly
I2C pull-ups (2x 4.7k): 1.4mA during communication
Powered-off sensor with leakage: 1-50uA
Pitfall 3: Underestimating Radio Energy
Wi-Fi transmission bursts may be short, but they dominate energy consumption:
Wi-Fi TX: 160-280mA for 100ms = 16-28 mAs per transmission
At 1 transmission/minute: adds 0.27-0.47 mA average current
This alone limits battery life to ~3-4 months on a 2000mAh battery
Pitfall 4: Not Testing at Temperature Extremes
Battery capacity drops dramatically in cold:
-20C: 65% of rated capacity
-40C: 30-40% of rated capacity
Always design for worst-case deployment temperature
2.12 Why Energy Efficiency Matters: Lessons from Biology
Nature has optimized for energy efficiency over billions of years. IoT designers can learn from biological systems:
Figure 2.3: Biological energy efficiency strategies and their IoT equivalents
Key biological principles applied to IoT:
Biological Strategy
IoT Implementation
Energy Savings
Hibernation
Deep sleep modes
1,000-10,000×
Torpor (short-term dormancy)
Light sleep between tasks
10-100×
Cold-blooded metabolism
Dynamic voltage/frequency scaling
2-8×
Event-driven response
Interrupt-based wake-up
Eliminates polling
Collective intelligence
Edge processing + cloud aggregation
Reduces communication
Worked Example: Battery Life Calculation for LoRaWAN Tracker
Scenario: GPS asset tracker with LoRaWAN reporting every 15 minutes.
Given:
Battery: 2,400 mAh @ 3.7V (Li-SOCl₂)
GPS fix: 40 mA for 30 seconds
LoRa TX (SF7): 100 mA for 1.5 seconds
MCU active: 10 mA for 2 seconds
Deep sleep: 5 µA
Energy per cycle (15 minutes = 900 seconds):
GPS: 40 mA × 30s = 1,200 mAs
LoRa: 100 mA × 1.5s = 150 mAs
MCU: 10 mA × 2s = 20 mAs
Sleep: 0.005 mA × 866.5s = 4.33 mAs
Total: 1,374.33 mAs per cycle
Cycles per day: 96
Daily energy: 96 × 1,374.33 = 131,935.68 mAs = 36.65 mAh
Battery life: 2,400 mAh ÷ 36.65 mAh = 65.5 days
With 70% efficiency (temp/self-discharge): 46 days
Optimization: Reduce GPS time using A-GPS (assisted GPS):
A-GPS: 10 seconds instead of 30 seconds
GPS energy: 40 × 10 = 400 mAs (saves 800 mAs per cycle)
New daily: 131,935.68 - (800 × 96) = 55,135.68 mAs = 15.32 mAh
New life: 2,400 ÷ 15.32 = 156.7 days → 110 days after derating
Result: 2.4× battery life improvement from A-GPS!
Putting Numbers to It
The battery gap quantifies why energy is IoT’s fundamental constraint:
Computing improvement (1990-2025): \(10^4 \times\) faster (Moore’s Law doubling every 18 months)
Battery improvement (1990-2025): \(10 \times\) energy density (7% annual improvement)
What this means: A 2025 IoT device has 1,000× more computing power per unit of battery capacity than a 1990 device. Energy efficiency MUST improve exponentially just to maintain the same battery life. This is why “sleep 99% of the time” is not optional—it’s the only way to bridge the gap.
Common Mistake: Assuming “Sleep Mode” Means Zero Power
The Mistake: Designer calculates battery life assuming sleep consumes 0 mA.
Reality: Even “deep sleep” consumes 1-10 µA for most MCUs, and peripheral leakage (voltage regulators, pull-ups, sensors) often adds 50-500 µA. For a 1% duty cycle device sleeping 99% of time, sleep current dominates total power consumption!
Fix: Measure actual sleep current with µA-capable meter. Target <10 µA total for multi-year battery life.
2.13 Knowledge Check
## Interactive Power Budget Calculator
Before diving deeper into energy concepts, try this interactive calculator to see how different design choices affect battery life:
Try These Scenarios
Ultra-low power sensor: Set duty cycle to 0.1%, sleep current to 5uA, active current to 20mA
Wi-Fi gateway: Set duty cycle to 50%, active current to 150mA (Wi-Fi)
LoRa sensor: Set duty cycle to 0.5%, active current to 40mA (LoRa TX)
Compare how dramatically battery life changes between these configurations!
2.14 Interactive Battery Life Calculator
Try adjusting the parameters below to see how different design choices affect battery life:
Show code
viewof battery_capacity = Inputs.range([100,10000], {value:2400,step:100,label:"Battery Capacity (mAh)"})viewof active_current = Inputs.range([1,300], {value:150,step:1,label:"Active Current (mA)"})viewof active_time = Inputs.range([0.1,60], {value:3,step:0.1,label:"Active Time per Cycle (seconds)"})viewof sleep_current = Inputs.range([0.001,1000], {value:10,step:0.001,label:"Sleep Current (µA)"})viewof cycle_interval = Inputs.range([1,1440], {value:5,step:1,label:"Cycle Interval (minutes)"})
Ultra-low power sensor: Set active current to 20mA, active time to 2s, sleep current to 5µA, cycle interval to 60 minutes
Wi-Fi camera: Set active current to 200mA, active time to 5s, sleep current to 100µA, cycle interval to 30 minutes
LoRa tracker: Set active current to 40mA, active time to 3s, sleep current to 10µA, cycle interval to 15 minutes
Notice how sleep current becomes dominant as you increase the cycle interval!
2.15 How It Works
Energy-aware IoT design operates on a fundamental principle: minimize active time, maximize sleep depth. Here’s the step-by-step process:
Energy Budget Calculation: Determine target battery life → Calculate maximum average current allowed → Work backwards to allocate energy per operation
State Machine Design: Define device states (deep sleep, sensor read, process, transmit) → Measure current draw for each → Calculate time in each state
Duty Cycle Optimization: Reduce wake frequency where possible → Batch operations to amortize overhead → Use event-driven wake (interrupts) instead of polling
Sleep Mode Selection: Match sleep depth to wake latency requirements → Use deepest sleep mode that still allows required wake sources (timer, GPIO, touch)
Real-world workflow: A temperature sensor reading every 10 minutes sleeps 99.9% of the time at 10µA, wakes for 2 seconds to read sensor (20mA) and transmit via LoRa (40mA), then returns to sleep. Average current: (40mA × 2s + 0.01mA × 598s) / 600s = 0.143mA, giving 5+ years on a CR123A battery.
2.16 Concept Relationships
Energy-aware design connects to multiple IoT domains:
Common dependency: Energy-aware design must be considered before finalizing hardware selection—choosing a 10µA vs 100µA sleep MCU has 10x battery life impact that software cannot fix.
Goal: Calculate battery life for different sleep configurations.
Given: ESP32 sensor, 2000mAh battery, reads sensor every 5 minutes - Active (sensor + WiFi TX): 150mA for 3 seconds - Light sleep: 800µA - Deep sleep: 10µA
Tasks:
Calculate average current with light sleep
Calculate average current with deep sleep
How much longer does deep sleep last?
What to observe: Deep sleep average ≈ 0.165mA (20+ years theoretical), light sleep ≈ 1.0mA (3 years). The 80× sleep current difference becomes only 6× battery life difference because TX dominates.
2.18.2 Exercise 2: Optimize a Real Design
Scenario: Wildlife camera, 6000mAh battery, 6-month target - Current: Active 200mA for 5s every 30min → 14 days battery life
Your task: Reduce average current from 0.93mA to 0.14mA (6 months)
Hints:
Can you reduce active time? (Better image compression, lower resolution)
Can you extend sleep interval? (Motion trigger instead of periodic)
Can you switch protocols? (BLE instead of WiFi for alerts)
What to observe: Changing interval 30min→2hr cuts daily cycles 48→12, saving 75% TX energy. Switching WiFi (160mA) to BLE (15mA) for alerts saves 90% TX power.
Matching Quiz: Match Energy Concepts to Definitions
Ordering Quiz: Order the Battery Life Estimation Process
Label the Diagram
💻 Code Challenge
Order the Steps
Match the Concepts
2.19 Summary
Energy-aware design is not optional in IoT—it’s fundamental. The key principles covered in this introduction:
Power vs Energy: Power is instantaneous consumption (W = V x I); energy is total consumption over time (Wh = P x t)
The Battery Gap: Battery technology improves slowly (5-8%/year) while computing advances rapidly, making energy efficiency increasingly critical
Design Tradeoffs: Every IoT design involves balancing edge vs. cloud processing, sleep depth vs. responsiveness, and complexity vs. power savings
