Industry Applications and Emerging Directions

Introduction

The Internet of Things (IoT) has developed from a collection of connected sensors into a foundation for digital transformation. Modern IoT systems combine sensing, communication, computation, artificial intelligence (AI), automation, and organisational decision-making. Their purpose is not simply to collect data, but to improve safety, quality, efficiency, sustainability, and human well-being.

An industrial IoT solution normally includes physical assets, sensors and actuators, local networks, edge computing, cloud or data-centre services, applications, and people. The most successful systems connect these layers to a clearly defined operational objective. For example, a manufacturing system may use vibration data to predict machine failure; a city may use traffic sensors to adapt signal timing; and a farm may use soil-moisture measurements to control irrigation.

IoT applications are therefore socio-technical systems. Technical performance is important, but so are work practices, regulation, cybersecurity, business value, skills, and public trust.

Learning Objectives

By the end of this chapter, the reader should be able to:

1. Explain IoT applications in manufacturing, cities, agriculture, healthcare, logistics, energy, and buildings.
2. Relate IoT to digital transformation and digital-twin architectures.
3. Describe autonomous systems, TinyML, edge AI, satellite IoT, and advanced mobile networks.
4. Evaluate benefits, limitations, security risks, ethical concerns, and sustainability implications.
5. Propose a practical IoT architecture for an industry problem.

1. From Connected Devices to Digital Transformation

Digital transformation is the organisational redesign of processes, services, and decisions using digital technologies. IoT contributes the connection between the physical and digital worlds. A sensor observes a physical condition, communication transports the observation, software interprets it, and an application supports or performs an action.

A useful transformation sequence is:

instrumentation → connectivity → visibility → intelligence → action → organisational change

Instrumentation provides measurements. Connectivity makes measurements available. Visibility presents conditions to operators. Intelligence identifies patterns or predicts events. Action changes a process, either through an actuator or a human decision. Organisational change occurs when the new capability becomes part of routine operations.

A dashboard alone is usually digitisation rather than transformation. Transformation occurs when the organisation changes how it schedules maintenance, manages inventory, delivers healthcare, or uses energy.

Value dimensions

IoT projects commonly create value through:

• Operational efficiency: reduced downtime, energy consumption, waste, or manual inspection.
• Quality improvement: earlier detection of defects and more consistent processes.
• Safety: remote monitoring, hazard detection, and reduced exposure to dangerous environments.
• Asset utilisation: better use of machines, vehicles, buildings, and infrastructure.
• New services: usage-based contracts, remote support, and outcome-based business models.
• Sustainability: measurement and reduction of emissions, water use, material loss, and unnecessary travel.

A business case should compare the total cost of ownership with measurable benefits. Total cost includes devices, installation, connectivity, software, integration, maintenance, cybersecurity, training, and eventual replacement. Benefits should not be assumed merely because data is available.

2. A General Industry IoT Architecture

Although architectures differ, an industry system can be analysed through six logical layers:

1. Physical layer: machines, vehicles, crops, patients, buildings, grids, and environmental conditions.
2. Sensing and actuation layer: sensors measure variables and actuators influence physical processes.
3. Connectivity layer: fieldbus systems, Ethernet, Wi-Fi, Bluetooth Low Energy, LPWAN, cellular networks, satellite links, or industrial wireless systems.
4. Edge and control layer: gateways and edge computers filter data, enforce rules, provide local control, and continue operating during cloud disconnection.
5. Data and platform layer: time-series databases, device management, identity services, digital models, analytics, and APIs.
6. Application and governance layer: dashboards, enterprise systems, workflows, human decisions, compliance, and accountability.

Protocols are selected according to range, power, latency, reliability, determinism, throughput, interoperability, and security. MQTT is widely used for lightweight publish–subscribe messaging; CoAP is suited to constrained web-like devices; OPC UA supports industrial information modelling and interoperability; HTTP is common for enterprise APIs; and 5G, Wi-Fi, Ethernet, LPWAN, and satellite systems provide different connectivity choices. No single protocol is optimal for every layer.

A sound architecture separates operational technology (OT), which controls physical processes, from information technology (IT), which manages enterprise data and applications. Controlled interfaces, identity management, network segmentation, and monitored gateways are essential between the two domains.

3. Manufacturing and Industrial IoT

Manufacturing IoT connects production equipment, robots, programmable logic controllers, quality systems, warehouse systems, and enterprise planning. Typical data includes temperature, pressure, current, vibration, position, cycle time, tool wear, and product measurements.

Predictive maintenance

Preventive maintenance replaces or services equipment according to a schedule. Predictive maintenance uses condition data to estimate when intervention is needed. A basic workflow is:

1. Acquire sensor and maintenance-history data.
2. Establish normal operating behaviour.
3. Detect abnormal patterns.
4. Estimate failure risk or remaining useful life.
5. Schedule maintenance and verify the result.

For a simple availability calculation, if a machine operates for time T and experiences downtime D:

Here, T and D must use the same time unit, and the formula assumes that downtime is correctly recorded and that the observation period represents normal operation. Availability alone does not measure product quality, safety, or maintenance cost.

Quality and flexible production

Machine-vision systems can inspect components, while connected tools record torque and process parameters. Digital work instructions can adapt to product variants. Robots and collaborative robots can support repetitive or hazardous tasks, but risk assessment must consider unexpected motion, sensor failure, and human interaction.

Industrial systems often require deterministic communication and high availability. A cloud-only design may be unsuitable for emergency shutdown or motion control because wide-area latency and service outages cannot be ignored. Local control should remain safe even when higher-level services fail.

4. Smart Cities and Infrastructure

Smart-city IoT includes transport, water, waste, public safety, lighting, environmental monitoring, and public services. Examples include adaptive traffic signals, parking occupancy systems, leakage monitoring, air-quality networks, connected street lighting, and flood warnings.

The city is a shared environment rather than a single organisation. Transport authorities, utilities, emergency services, private operators, residents, and regulators may own different data and infrastructure. Interoperability and data governance are therefore as important as sensors.

A city should avoid deploying surveillance systems merely because technology permits it. Data minimisation, purpose limitation, retention controls, public communication, and independent oversight are necessary, especially when cameras, location data, biometrics, or predictive analytics are involved. Smart-city benefits should be distributed fairly; otherwise, connected services may increase inequality between neighbourhoods.

5. Agriculture and Environmental Monitoring

Precision agriculture applies IoT to soil, crops, livestock, weather, irrigation, and machinery. Sensors may measure soil moisture, electrical conductivity, temperature, humidity, leaf conditions, or animal movement. Satellite imagery and drones can complement ground measurements.

A simple irrigation decision may compare measured volumetric soil-water content θ with a lower threshold θ_min. Irrigation is considered when:

This is a control rule, not a universal agronomic law. Thresholds depend on crop type, soil properties, root depth, weather, and growth stage. Poorly calibrated sensors can cause over-irrigation or crop stress.

IoT can reduce water use by applying irrigation by zone and responding to actual conditions. However, rural deployments face long distances, weak connectivity, battery replacement difficulties, harsh weather, and limited technical support. Low-power wide-area networks and satellite links can extend coverage, while local storage allows operation during communication outages.

6. Healthcare and Assisted Living

Healthcare IoT includes wearable devices, remote patient monitoring, smart hospital equipment, medication systems, asset tracking, and environmental controls. Devices may measure heart rate, oxygen saturation, glucose, temperature, movement, or blood pressure.

Healthcare systems require a higher standard of safety and evidence than ordinary consumer applications. A device that detects an abnormal reading must define who receives the alert, how quickly it is reviewed, and what action is authorised. An algorithmic alert without clinical workflow may create alarm fatigue rather than better care.

Important requirements include:

• informed consent and understandable privacy notices;
• data integrity and accurate timestamps;
• secure identity and access control;
• clinical validation and documented limitations;
• interoperability with electronic health records;
• fail-safe behaviour and human clinical responsibility;
• compliance with applicable medical-device and data-protection law.

IoT can support ageing in place, but monitoring should preserve dignity and autonomy. Continuous monitoring must not become unjustified surveillance.

7. Logistics, Fleet Operations, and Supply Chains

Logistics IoT tracks vehicles, containers, pallets, packages, and environmental conditions. GPS, cellular communication, RFID, Bluetooth beacons, temperature sensors, and telematics are commonly combined.

A cold-chain system may record temperature throughout transport. If a medicine must remain within [MATH:T_min] and [MATH:T_max], an excursion occurs when:

The system must record not only the temperature but also time, location, sensor identity, calibration status, and chain of custody. A single measurement without context may not support a valid quality decision.

IoT improves estimated arrival times, route planning, asset utilisation, theft detection, and proof of delivery. It also introduces risks: location data can reveal employee behaviour, trackers can be tampered with, and inaccurate data can cause costly routing decisions. Secure device mounting, tamper evidence, authenticated updates, and exception-based human review are practical controls.

8. Energy Systems and Smart Grids

Energy IoT connects meters, distributed generation, batteries, substations, buildings, electric vehicles, and control centres. Smart meters provide consumption data; sensors monitor equipment; controllers coordinate demand response and storage.

A basic energy relationship is:

where E is energy, P is power, and t is time. If P is measured in kilowatts and t in hours, E is measured in kilowatt-hours. The relationship assumes constant power. For varying power, energy is accumulated over time:

Smart grids must balance reliability, affordability, cybersecurity, and decarbonisation. Distributed energy resources can improve resilience but also increase system complexity. A compromised inverter, meter, or charging station may affect a large number of assets. Grid operators therefore need strong authentication, segmentation, monitoring, recovery procedures, and secure supply chains.

9. Smart Buildings

Building IoT integrates heating, ventilation and air conditioning, lighting, access control, elevators, occupancy sensing, fire systems, and energy meters. Building-management systems can adjust operation according to occupancy, indoor air quality, weather, and electricity prices.

The most effective approach is usually measured optimisation rather than indiscriminate automation. Occupancy sensors may reduce energy consumption, but poorly tuned controls can reduce comfort or create unsafe ventilation conditions. Building owners should establish baseline energy performance, define comfort limits, test control changes, and verify results.

Legacy building systems often remain in service for decades. Retrofitting secure gateways and segmenting networks may be more practical than replacing every controller. Cybersecurity responsibilities must be clear among owners, facility managers, integrators, and service providers.

10. Digital Twins

A digital twin is a maintained digital representation of a physical asset, process, or system, linked through data and used for observation, analysis, simulation, or control. A static computer-aided design model is not necessarily a digital twin because it may lack operational data and lifecycle updating.

A practical twin includes:

• an identity for the physical asset;
• a model of structure, state, and relationships;
• time-stamped observations;
• links to maintenance and operational history;
• simulation or analytical functions;
• rules for synchronisation and data quality;
• access control and lifecycle governance.

Digital twins range from a component-level pump model to a city-scale infrastructure model. Their accuracy must match the decision. A high-fidelity model is not automatically valuable if its data is late, incomplete, or unauthorised for the intended use.

Twins support commissioning, training, predictive maintenance, what-if analysis, and lifecycle planning. They also create a risk of false confidence: model outputs are estimates based on assumptions and should be validated against physical evidence.

11. Autonomous Systems and Human Oversight

An autonomous system senses its environment, interprets conditions, selects actions, and acts with limited direct human control. Examples include mobile robots, automated warehouses, drones, autonomous vehicles, and adaptive process controllers.

Autonomy should be described by capability and operating domain. A warehouse robot operating on mapped floors in controlled conditions is not equivalent to a vehicle operating on public roads in all weather. Designers should specify the operational design domain, prohibited conditions, fallback behaviour, and handover requirements.

A safe control loop is:

sense → estimate → decide → act → verify

Verification is critical. The system should detect sensor disagreement, communication loss, unsafe commands, and conditions outside its training or design domain. Human oversight must be meaningful, with sufficient time, information, authority, and competence to intervene.

12. TinyML and Edge AI

TinyML places machine-learning inference on microcontrollers or other highly constrained devices. Edge AI performs inference near the data source, such as on a gateway, camera, robot, or industrial computer.

Benefits include low latency, reduced bandwidth, improved privacy, resilience during cloud outages, and lower communication cost. Constraints include limited memory, processing power, energy, model size, training data, and update capability.

A typical workflow is:

1. Define the event or decision precisely.
2. Collect representative and ethically obtained data.
3. Label data and examine bias and quality.
4. Train and validate a model.
5. Compress or quantise it for the target hardware.
6. Test accuracy, latency, energy, and failure behaviour.
7. Deploy securely and monitor drift.

Inference accuracy is not the only metric. False negatives may be more dangerous than false positives in safety applications, while excessive false positives may cause alarm fatigue. Models must be monitored because equipment, populations, environments, and behaviours change.

13. Satellite IoT, 5G, and 6G Directions

Satellite IoT is valuable where terrestrial networks are absent or unreliable, including oceans, remote farms, pipelines, mines, and disaster zones. Its disadvantages may include higher latency, limited bandwidth, terminal cost, atmospheric or obstruction effects, and power constraints. Hybrid systems can use terrestrial links when available and satellite communication as a fallback.

5G supports IoT through improved capacity, mobility, device density, and potential low-latency services. However, advertised capability is not identical to guaranteed application performance. Industrial users must evaluate coverage, spectrum, network configuration, service-level agreements, security, and integration with OT.

6G remains an emerging research and standardisation direction rather than a universally deployed technology. Proposed themes include integrated sensing and communication, AI-native networking, highly distributed edge computing, non-terrestrial integration, and more efficient use of spectrum and energy. Such claims should be distinguished from established operational capabilities.

14. Sustainable and Resilient IoT Architectures

IoT itself consumes materials and energy through devices, networks, data centres, batteries, and maintenance travel. A sustainable architecture considers the entire lifecycle:

• use only necessary sensors and sampling rates;
• select energy-efficient protocols and processors;
• design for repair, secure updates, and long service life;
• reduce unnecessary data transmission and storage;
• measure embodied carbon and electronic waste;
• use renewable energy where practical;
• plan responsible reuse and disposal.

Resilience means continuing essential functions during faults, attacks, power loss, or connectivity disruption. Edge autonomy, local control, redundant sensing, graceful degradation, backups, and tested recovery plans support resilience.

Security should be designed from the beginning. Fundamental practices include unique device identities, secure boot, encryption in transit and at rest, least privilege, network segmentation, vulnerability management, signed updates, logging, time synchronisation, and retirement procedures. A device that cannot be updated securely may become a long-term liability.

15. Illustrative Case: Connected Cold-Chain Logistics

Background

A regional pharmaceutical distributor transports temperature-sensitive medicines from a warehouse to clinics. The organisation experiences occasional temperature excursions and cannot always determine where they occurred.

Problem

The existing system records temperature manually at dispatch and delivery. It lacks continuous evidence, calibrated sensors, reliable alerts, and clear responsibility for responding to exceptions.

Analysis

The critical requirements are measurement accuracy, continuous logging, location context, alert delivery, battery life, tamper resistance, and regulatory traceability. A cloud dashboard alone would not solve the problem if drivers cannot receive alerts or if data is missing during network outages.

Approach and solution

Each container receives a calibrated temperature sensor with a unique identity. The sensor stores readings locally and transmits through a gateway using cellular or short-range communication. The gateway forwards data to a cloud service when coverage is available. A rules engine compares readings with the permitted range, records excursions, and sends an alert to the responsible logistics coordinator.

The architecture includes:

• signed device firmware and authenticated gateways;
• local buffering during network loss;
• time synchronisation and calibration records;
• role-based access for drivers, quality staff, and managers;
• an audit trail linking shipment, container, sensor, route, and receiver;
• a documented response procedure for excursions.

Outcome

The organisation gains continuous visibility and can distinguish a brief sensor anomaly from a sustained excursion. It can identify recurring problems such as loading-bay delays or inadequate vehicle insulation. The system does not automatically declare medicine safe or unsafe; qualified personnel interpret evidence according to approved procedures.

Learning points

The case demonstrates that IoT value depends on workflow, calibration, governance, and response—not merely on installing sensors. Local storage and clear escalation are essential for resilience. Data quality and chain of custody determine whether measurements can support professional decisions.

16. Challenges, Ethics, and Governance

Important challenges include fragmented standards, legacy equipment, unreliable connectivity, cybersecurity exposure, poor data quality, vendor lock-in, skills shortages, and uncertain return on investment. Large deployments also create operational complexity: thousands of devices require inventory, monitoring, patching, replacement, and secure decommissioning.

Ethical issues include privacy, consent, surveillance, discrimination, accessibility, labour impacts, and accountability for automated decisions. Data collected for one purpose should not be silently reused for another. AI-based decisions should be explainable enough for affected people and operators to understand relevant limitations.

Governance should assign ownership for data, devices, models, incidents, and decisions. Organisations should maintain an asset register, risk assessments, retention rules, access policies, incident procedures, supplier requirements, and audit evidence. Compliance requirements differ by sector and jurisdiction; technical teams must consult applicable telecommunications, safety, privacy, cybersecurity, medical-device, environmental, and critical-infrastructure regulations.

17. Best-Practice Design Process

A practical IoT project can follow these stages:

1. Define the operational problem and success measures.
2. Identify stakeholders, hazards, legal duties, and affected communities.
3. Map assets, data flows, protocols, dependencies, and failure modes.
4. Select the minimum viable sensing and control architecture.
5. Build a small pilot using representative conditions.
6. Test accuracy, latency, availability, security, usability, and total cost.
7. Integrate with maintenance, enterprise, and emergency workflows.
8. Deploy gradually with monitoring and rollback capability.
9. Review benefits, incidents, bias, energy use, and lifecycle costs.
10. Retire devices and data securely when the system is no longer justified.

Useful indicators may include availability, alert precision, response time, energy per device, battery life, avoided downtime, water saved, carbon reduction, safety incidents, and percentage of devices receiving security updates. Metrics should reflect outcomes rather than vanity measures such as device count.

Chapter Summary

IoT has become an enabling architecture for digital transformation across manufacturing, smart cities, agriculture, healthcare, logistics, energy, and buildings. Its value arises when physical observations are converted into reliable decisions and actions within an operational workflow. Effective architectures combine sensing, suitable protocols, edge processing, platforms, applications, governance, and cybersecurity.

Digital twins extend IoT by maintaining operational representations of physical assets and processes. Autonomous systems add decision and action capabilities, while TinyML and edge AI move intelligence closer to data sources. Satellite IoT extends coverage to remote locations, and 5G and future 6G directions offer new possibilities for connectivity and distributed intelligence. These technologies must be evaluated realistically rather than by marketing claims.

Sustainable IoT requires lifecycle thinking, efficient data practices, repairability, secure updates, resilience, and responsible disposal. Privacy, safety, fairness, accountability, and human oversight are central design requirements. The strongest professional approach begins with a well-defined problem, validates the business and social value, pilots safely, and governs the system throughout its life.

Key Terms

Digital transformation: Organisational improvement through digitally enabled processes, services, and decisions.

Industrial IoT: IoT applied to industrial assets, production, maintenance, and operations.

Operational technology: Hardware and software that monitors or controls physical processes.

Edge computing: Processing data near the source rather than exclusively in a remote cloud.

TinyML: Machine learning deployed on highly constrained devices such as microcontrollers.

Edge AI: Artificial-intelligence inference performed near the data source.

Digital twin: A maintained digital representation of a physical asset, process, or system.

Predictive maintenance: Maintenance planned using observed condition and predicted failure risk.

Smart city: An urban environment using connected data and technology to improve services and outcomes.

Precision agriculture: Data-guided management of crops, soil, livestock, and resources.

Cold chain: Controlled-temperature storage and transport for sensitive products.

Smart grid: An energy network using communication, sensing, and control to coordinate supply and demand.

Autonomous system: A system that senses, decides, and acts with limited direct human control.

LPWAN: Low-power wide-area network designed for long-range, low-data-rate devices.

Interoperability: The ability of systems to exchange and correctly use information.

Network segmentation: Dividing a network into controlled zones to limit threats and failures.

Data minimisation: Collecting and retaining only data necessary for a defined purpose.

Graceful degradation: Continued safe operation at reduced capability during failures.

Total cost of ownership: All acquisition, operation, maintenance, security, and retirement costs.

Review Questions

Short Answer Questions

1. What distinguishes digital transformation from simple data collection?
2. State two reasons why edge computing is useful in IoT systems.
3. What is a digital twin?
4. Identify two ethical concerns in smart-city IoT.
5. Why is local buffering useful in a cold-chain system?

Long Answer Questions

1. Explain the six logical layers of a general industry IoT architecture and describe their relationships.
2. Compare IoT applications in manufacturing, agriculture, and healthcare.
3. Discuss the benefits and limitations of predictive maintenance.
4. Explain how TinyML and edge AI affect IoT architecture and operations.
5. Describe the security and sustainability requirements of a long-lived IoT deployment.

Analytical / Application Questions

1. A rural irrigation project has intermittent connectivity, battery-powered sensors, and several soil types. Propose an architecture and explain how it would make irrigation decisions safely.
2. Design a connected-building pilot to reduce energy use without reducing ventilation or occupant comfort. Specify measurements, controls, risks, and success indicators.
3. A factory wants to place a cloud-controlled robot on a production line. Analyse whether cloud-only control is appropriate and recommend a safer architecture with fallback behaviour.

Multiple Choice Questions with Answers

Question 1

Which sequence best represents the progression from sensing to organisational transformation?

A. Action → sensing → connectivity → visibility → intelligence

B. Instrumentation → connectivity → visibility → intelligence → action

C. Connectivity → disposal → sensing → action → visibility

D. Intelligence → instrumentation → storage → connectivity → action

Correct Answer: B. Instrumentation → connectivity → visibility → intelligence → action

Question 2

Which task is most appropriately retained at the edge in a safety-critical industrial system?

A. Long-term business reporting

B. Annual equipment budgeting

C. Emergency local shutdown control

D. Archiving historical maintenance documents

Correct Answer: C. Emergency local shutdown control

Question 3

A machine operates for 100 hours and is unavailable for 8 hours. Using Availability = (T − D) / T × 100%, what is its availability?

A. 8%

B. 92%

C. 100%

D. 108%

Correct Answer: B. 92%

Question 4

Which statement best defines a digital twin?

A. A static product photograph

B. A maintained digital representation linked to a physical asset or process

C. A replacement for all physical sensors

D. A network cable connecting two identical devices

Correct Answer: B. A maintained digital representation linked to a physical asset or process

Question 5

Which condition represents a temperature excursion when the permitted range is T_min to T_max?

A. T_min ≤ T ≤ T_max

B. T = 0 only

C. T < T_min or T > T_max

D. T is recorded with a timestamp

Correct Answer: C. T < T_min or T > T_max

Question 6

What is a principal benefit of TinyML?

A. It requires unlimited cloud bandwidth

B. It enables inference on highly constrained devices

C. It eliminates the need for representative data

D. It guarantees that models never drift

Correct Answer: B. It enables inference on highly constrained devices

Question 7

Which practice most directly supports IoT sustainability?

A. Collecting every possible measurement indefinitely

B. Replacing devices at every software update

C. Designing for repair, secure updates, and long service life

D. Avoiding all local processing

Correct Answer: C. Designing for repair, secure updates, and long service life

Question 8

Why is calibration important in a healthcare or cold-chain IoT system?

A. It increases the colour quality of dashboards

B. It helps ensure measurements support valid decisions

C. It removes the need for cybersecurity

D. It guarantees uninterrupted network coverage

Correct Answer: B. It helps ensure measurements support valid decisions

Question 9

Which characteristic is particularly associated with satellite IoT?

A. Useful coverage in locations lacking terrestrial networks

B. Zero latency in all conditions

C. Unlimited bandwidth at no cost

D. No requirement for an antenna or terminal

Correct Answer: A. Useful coverage in locations lacking terrestrial networks

Question 10

Which control best limits the impact of a compromised IoT device?

A. Network segmentation and least-privilege access

B. Using the same password on every device

C. Disabling all software updates

D. Connecting every device to one unrestricted network

Correct Answer: A. Network segmentation and least-privilege access