Unit 6: Embedded system and its application

An embedded system is a dedicated computing unit built into a larger product to perform a fixed set of tasks, as opposed to a general-purpose computer. The field matured alongside microcontrollers in the 1970s (Intel 8048, 1976) and now dominates device counts worldwide. Everything in this unit rests on a few defining traits.

• Dedicated function: Runs one application defined at design time; a washing-machine controller never becomes a word processor.
• Real-time constraint: Must respond within a deadline; an airbag controller firing in milliseconds is hard real-time, a thermostat is soft real-time.
• Resource-constrained: Limited RAM, flash and clock speed to minimise cost and power — measured in KB and MHz, not GB and GHz.
• Hardware–software co-design: Firmware is written for one fixed board; software and silicon are tuned together.
• Reliability and autonomy: Operates unattended, often with a watchdog timer that resets the system if the program hangs.

II. Introduction to Embedded System

The dedicated computer hidden inside a product

An embedded system combines a processing core, memory and peripherals to sense, decide and actuate within a host device.

A. Definition and building blocks

It is a microprocessor- or microcontroller-based system committed to a specific control task.

• Processing core: The CPU (microcontroller/microprocessor/DSP) executing firmware from non-volatile memory.
• Memory: Flash/ROM holds the program; RAM holds runtime variables; the split is fixed and small.
• Input peripherals: Sensors and switches feeding the core via ADC, GPIO or serial lines.
• Output peripherals: Actuators, displays and relays driven through PWM, DAC or GPIO.
• Firmware: The embedded program, often running on a bare-metal superloop or a real-time OS (RTOS) such as FreeRTOS.

B. Classification and significance

Embedded systems scale from trivial to safety-critical.

• By performance: Small-scale (8-bit, e.g. remote control), medium-scale (16/32-bit, e.g. router), sophisticated (multicore + RTOS, e.g. avionics).
• By timing: Real-time (anti-lock brakes) versus standalone (digital camera).
• Significance: Low unit cost, low power (µA sleep modes), compact size and deterministic behaviour make them the backbone of consumer, industrial and automotive products.

III. Comparison of Microprocessor and Microcontroller

General-purpose CPU versus system-on-chip

Both execute instructions, but they differ in integration, which decides where each is used.

A. Microprocessor

A microprocessor is a single-chip CPU that needs external support chips to form a usable system.

• Architecture: Typically Von Neumann — program and data share one memory bus (e.g. Intel 8085, 8086).
• External dependency: RAM, ROM, timers and I/O ports are added as separate ICs on the board.
• Strength: High clock speeds and large addressable memory suit computation-heavy, flexible tasks.

B. Microcontroller

A microcontroller integrates CPU, memory and peripherals on one chip for embedded control.

• Architecture: Often Harvard — separate program and data buses allowing fetch and execute in parallel (e.g. 8051, AVR, PIC).
• On-chip peripherals: ADC, timers, UART, SPI, I²C and GPIO built in, cutting board size and cost.
• Strength: Bit-level instructions, low power and deterministic timing suit dedicated control.

C. Point-by-point comparison

The contrast guides device selection.

1. Microprocessor: External memory and I/O; higher power and cost per system; best for computing (PCs, servers).
2. Microcontroller: Self-contained; lower power and cost; best for control (appliances, instruments).

IV. Introduction of Arduino and Sensors

An open prototyping platform and the devices that feed it data

Arduino is an open-source microcontroller board and IDE that lowers the barrier to embedded development; sensors are the transducers that convert physical quantities into electrical signals it can read.

A. Introduction of Arduino

Arduino (Italy, 2005) is a family of boards built around Atmel AVR microcontrollers with a simple C/C++ toolchain.

• Hardware: The Uno uses an ATmega328P — 32 KB flash, 2 KB RAM, 16 MHz clock, 14 digital pins and 6 analog inputs (10-bit ADC).
• Programming model: Two mandatory functions —
  
  CPP

    void setup() { pinMode(13, OUTPUT); }     // runs once
    void loop()  { digitalWrite(13, HIGH); }  // repeats forever

• Interfaces: Built-in UART, SPI, I²C and PWM pins let it talk to sensors, displays and modules without extra silicon.
• Ecosystem: Shields (stackable boards) and libraries abstract complex hardware into one-line calls, e.g. analogRead(A0).

B. Sensors

A sensor converts a physical parameter into a measurable electrical output for the controller.

• Analog sensors: Produce continuous voltage read by the ADC — e.g. LM35 giving 10 mV/°C, so 250 mV = 25 °C.
• Digital sensors: Output coded data over a bus — e.g. DHT11 (temperature/humidity) over a single-wire protocol.
• Common types: IR/ultrasonic for distance, PIR for motion, LDR for light, MPU-6050 accelerometer/gyroscope for motion sensing.
• Signal conditioning: Amplification, filtering and ADC conversion turn weak raw signals into clean digital values; a 10-bit ADC maps 0–5 V to 0–1023 counts.

V. Role of IoT and Cloud Computing in Applications

Connecting embedded sensors to networked intelligence

The Internet of Things (IoT) links embedded devices to networks so their data can be gathered, stored and analysed; cloud computing supplies the remote storage and processing that scales this. Condition monitoring — tracking the health of equipment through sensor data to predict faults before failure — is the thread running through every domain below.

A. Condition monitoring of plant processes

IoT turns industrial plant monitoring from periodic manual checks into continuous remote supervision.

• Sensing: Vibration, temperature and pressure sensors on pumps and motors stream data via gateways to the cloud.
• Cloud role: Stores time-series history and runs analytics to flag anomalies — e.g. a rising vibration RMS predicting bearing wear.
• Benefit: Predictive maintenance replaces reactive repair, cutting unplanned downtime.

B. Health care

Connected medical devices extend monitoring beyond the hospital bed.

• Sensing: Wearables track heart rate, SpO₂ (pulse oximetry) and glucose; data uploads over Bluetooth/Wi-Fi.
• Cloud role: Aggregates patient records, triggers alerts to clinicians, and enables remote patient monitoring and telemedicine.
• Benefit: Early warning of deterioration and continuous care for chronic patients.

C. Agriculture

IoT enables precision farming that applies inputs only where needed.

• Sensing: Soil-moisture, pH, temperature and humidity sensors across a field.
• Cloud role: Combines sensor data with weather forecasts to schedule irrigation and fertiliser automatically.
• Benefit: Water and input savings with higher yield; pumps actuate only when soil moisture drops below a set threshold.

D. Manufacturing

The Industrial IoT (Industry 4.0) links machines, products and systems on the factory floor.

• Sensing: Machine controllers and RFID tags report throughput, energy use and tool status.
• Cloud role: Builds a digital twin — a live virtual model — to optimise scheduling and detect quality drift.
• Benefit: Higher Overall Equipment Effectiveness (OEE) and traceability across the line.

E. Automobiles

Connected vehicles use embedded sensors plus cloud links for safety and efficiency.

• Sensing: Engine, GPS, radar and camera data from on-board electronic control units (ECUs).
• Cloud role: Enables fleet telematics, over-the-air (OTA) firmware updates and vehicle-to-everything (V2X) coordination.
• Benefit: Predictive servicing, navigation and foundations for advanced driver assistance.

F. Smart grid

IoT converts the one-way power grid into a two-way, self-monitoring network.

• Sensing: Smart meters and feeder sensors report consumption and line status in real time.
• Cloud role: Balances supply and demand, integrates solar/wind inputs, and isolates faults automatically.
• Benefit: Demand-response pricing, reduced losses and faster outage recovery.

G. Common architecture and limitations

A shared four-layer stack underlies every application above.

• Layered model:
  • Perception layer: Sensors and microcontrollers capture data.
  • Network layer: Wi-Fi, Zigbee, LoRa or cellular transport it.
  • Processing layer: Cloud/edge servers store and analyse.
  • Application layer: Dashboards and alerts deliver insight to users.
• Edge computing: Processing data near the sensor cuts latency and bandwidth for time-critical decisions.
• Limitations: Security and privacy of streamed data, dependence on connectivity, and power budgets for battery devices remain the central design challenges.