1 INTERNET OF THINGS(IOT) SENSORS:

2 STATE-OF-THE-ART AND FUTURE

3 PROSPECTS
4 Anshu Abhishek Raj 14 24
5 Electrical and Computer 15 A Venkata Durga
6 Engineering Department of 16 Electrical and Computer 25 Amruthesh M N
7 Electrical and Electronics 17 Engineering Department of 26 Electrical and Computer
8 engineering 18 Electrical and Electronics 27 Engineering Department of
9 Amrita Vishwa Vidyapeetham, 19 engineering 28 Electrical and Electronics
10 Kollam, Kerala, India 20 Amrita Vishwa Vidyapeetham, 29 engineering
11 am.en.u4elc22009@am.students.a 21 Kollam, Kerala, India 30 Amrita Vishwa Vidyapeetham,
12 mrita.edu 22 am.en.u4elc22007@am.students.a 31 Kollam, Kerala, India
13 23 mrita.edu 32 am.en.u4elc22008@am.students.a
33 mrita.edu

34 Abstract — The IoT relies on interconnected devices, 72 modules. We delve into the realm of virtual and smart
35 with sensors as its senses. This review explores sensor73 sensors, where virtual entities glean insights without a
36 technology, from traditional to virtual and smart 74 physical presence, and smart sensors adapt their
37 sensors. We examine sensor data processing, including
75 functionalities based on environmental cues. But the
38 fusion and analysis, and IoT applications integrating AI,
39 in smart cities, homes, Industry 4.0, and agriculture.
76 journey doesn't end there. We then dissect the
40 Sensor types are classified based on principles, 77 intricate architecture of IoT sensor data processing.
41 highlighting their data-capturing role. Virtual and 78 Imagine raw data, a cacophony of numbers and
42 smart sensors offer insights without physical form. IoT79 readings, flowing from countless sensors. This is
43 processing involves fusion, aggregation, 80
and where the magic happens. Through fusion, disparate
44 transformation for meaningful data. IoT and AI enable 81 data streams are merged, revealing hidden patterns
45 automation, predictive maintenance, and smart 82 and correlations. Sophisticated analysis techniques
46 decisions. Smart cities optimize traffic, manage energy,
83 extract meaningful information, identifying
47 and enhance safety. Challenges include security,
84 anomalies and predicting future trends. Finally,
48 privacy, and standardization. Innovations include
49 miniaturization, energy efficiency, and self-healing, 85 transformation sculpts the data into actionable
50 promising a smarter, sustainable future. 86 insights, ready to drive intelligent decisions and
87 automated actions. Empowered with this
51 Keyword: Internet of Things (IoT), Smart Sensors,
88 comprehensive understanding, we embark on a
52 Virtual sensors, Artificial intelligence 89 captivating exploration of transformative applications
53 I. INTRODUCTION 90 fueled by the data gleaned from these sensors. We
91 witness the synergy between IoT and Artificial
54 Imagine a world where a cup of coffee whispers 92 Intelligence (AI), where sensor data serves as the
55 "brewed to perfection," your home anticipates your 93 lifeblood of AI algorithms, enabling smarter decision-
56 arrival by adjusting the thermostat, and fields 94 making in countless domains. We envision smart
57 communicate their optimal irrigation needs. This 95 cities where traffic flows intelligently, energy
58 symphony of interconnected objects, seamlessly 96 consumption is optimized, and citizen safety is
59 collecting and exchanging data, is the essence of the 97 enhanced – all guided by the intricate network of
60 Internet of Things (IoT). But the true magic lies not 98 interconnected sensors. Stepping into the smart home,
61 99
in the devices themselves, but in their silent partners: we experience a symphony of comfort and
62 100
sensors. Like the beating heart of the IoT ecosystem, convenience. Sensors anticipate our needs, adjusting
63 these miniature marvels act as our eyes and ears, 101 lighting, regulating temperature, and even monitoring
64 capturing the nuances of our physical and 102 our health and well-being. Industry 4.0 comes alive as
65 environmental surroundings. This review delves deep 103 sensor-driven automation transforms manufacturing,
66 into the fascinating world of IoT sensor technology,104 optimizing production lines, predicting equipment
67 unveiling its profound impact on shaping a smarter 105 failures, and enabling real-time data-driven decision-
68 future. We begin by exploring the diverse tapestry106 of making. The agricultural landscape undergoes a
69 107
sensor types, ranging from traditional workhorses like revolution with the rise of smart agriculture. Sensors
70 temperature and pressure sensors to cutting-edge 108 monitor soil conditions, track crop health, and
71 marvels like biometric sensors and image recognition109 optimize irrigation, leading to increased yields and
110 reduced environmental impact. From remote 151 A. Communication and working of Sensors in IoT
111 healthcare and personalized medicine to intelligent
112 transportation and environmental monitoring, 152 the For resource-constrained devices, RFID offers a fast-
113 applications of IoT sensor technology seem 153 growing solution for network communication and
114 154
boundless. However, navigating this transformative automated identification, unlocking new possibilities
115 journey is not without its challenges. Security 155 in this field. These RFID tags are integrated circuits
116 concerns arise as countless devices collect and 156 having an antenna. An RFID reader, through the
117 transmit data, demanding robust cybersecurity 157 radio interface, communicates with these RFID tags.
118 measures. Data privacy issues require careful 158 Fig. 1 shows how IoT works using various
119 159
consideration, ensuring responsible data collection technologies including RFID that identifies and tracks
120 and usage. The need for standardized protocols 160 data of things, sensors play the role of collecting and
121 becomes paramount to ensure 161
seamless processing of data for detecting changes in physical
122 interoperability and communication across diverse162 things, smart-tech enhances network power by
123 devices and networks. Yet, the future brims with 163 developing the capabilities in network and nanotech
124 immense promise. We look forward to cutting-edge 164 that enable smaller things to connect and interact with
125 sensor innovations, encompassing miniaturization,165 other things in the network. Communication with
126 enhanced energy efficiency, and self-healing 166 such devices can be done directly, hands-free,
127 capabilities. The rise of edge computing promises167 remotely over the Internet. This way of
128 localized data processing, reducing latency and 168 communication is known as an intelligent
129 enhancing security. As we move forward, embracing169 communication system.
130 the challenges and celebrating the innovations, 170
we Think of the IoT as a giant brain that connects
131 pave the way for a smarter, more efficient, and 171 devices and collects data, opening up a world of
132 sustainable future. This future is powered by 172the possibilities. IoT services differ from other IT-related
133 sensory web, a tapestry woven from countless 173 techniques found in industrial or home office context
134 interconnected sensors, whispering insights that will
174 due to their ubiquitous and embedded characteristics
135 redefine our world. This review serves as 175 an that permeate our daily lives. While the main IoT
136 invitation to explore this captivating realm, 176 architecture(s) is still under definition, their
137 understanding how its technology paves the way for a
177 underlying paradigm offers the means to gather one
138 future shaped by intelligence, automation, and 178 or more “things” (sensors, interactive devices, or
139 human-centric solutions. 179 even complex ones) using well-defined
140 II. IOT AND SENSORS 180 communication interfaces. These can then share data
181 and communicate with the outside world through
141 182
In the technology era of the Internet of Things (IoT), some specially designed network gateways. The final
142 the interconnection of physical things with virtual183 result is similar to service-oriented architectures that
143 objects aims to enhance human life quality through 184 can be subject to service orchestration (e.g., through
144 advanced applications and growing sensor 185 their APIs). Millions of sensors embedded in IoT
145 technology, communication networks and processing 186 devices constantly gather data, creating a big data
146 methodologies. The wireless-based sensor network187 is deluge. Frameworks analyse this data from sensors
147 much more popular, as these networks can 188 be in:
148 deployed ad hoc without the prerequisite of any 189 Smart buildings: Managing oxygen levels, detecting
149 190
infrastructure. The wireless sensor network is capable smoke/hazardous gases, and measuring luminosity.
150 of self-organizing and can be deployed randomly. 191 Industrial production: Optimizing processes and
192 improving efficiency.
193 Fig.1 The basic architecture of IoT sensor data processing, data fusion and data analysis
194 248
Traffic management: Collecting real-time information method allows for a high level of inferences of
195 for congestion control. 249 knowledge and much accurate decision-making. In
196 Healthcare: Monitoring patient health conditions250
in recent years, the adoption of emerging technologies
197 real-time. 251 has revolutionized cloud computing, fog computing
198 Manufacturing: Transforming traditional approaches
252 and edge computing towards IoT sensor data analysis.
199 with data-driven insights. 253 These enabling technologies provide a pervasive,
200 This sensor data, analysed through big data and deep
254 reliable and convenient platform to handle IoT sensor
201 learning, fuels a high-tech society with:
255 data’s dynamic, heterogeneous nature. As such, the
202 1. Improved decision-making: Across various 256 data analytic layer aims at developing smart
203 sectors, from smart cities to healthcare. 257 functionality to address a wide variety of IoT-based
204 258
2. Enhanced efficiency: In areas like industrial applications. The objectives of these platforms are to
205 production and energy management. 259 reduce the computation and storage cost, improve
206 3. Real-time insights: Enabling proactive 260 network transmission reliability, reduce the network
207 actions and better outcomes. 261 delay, enhance IoT network security and privacy,
262 ensure scalability, and allow failure- and risk-free IoT
208 The wireless sensor network is capable of self- 263 solutions. Sensor data is classified according to its
209 organizing and can be deployed randomly. Figure264 1 application. It plays an important part in the
210 depicts the basic architecture for IoT sensor data 265 automation of any application by measuring and
211 processing, fusion and analysis layers. The IoT sensor
266 processing the collected data for detecting changes in
212 data layer primarily consists of various IoT sensors
267 physical things. Whenever there is a change in any
213 that can measure physical surroundings and capture 268 physical condition for which a sensor is made, it
214 real-time environment changes. The IoT sensors are
215 mainly associated with the micro processing unit,
216 storage unit, control unit, power system and wireless
217 communication interfaces. The IoT sensor devices are
218 constrained in size, computing power, memory,
219 networking capability and storage space. Wireless
220 communication protocols, such as Wi-Fi, Zig Bee,
221 Bluetooth, Near Frequency Communication (NFC)
222 and LTE/4G mobile technologies, are commonly 269 produces a measurable response. Fig. 2 shows
223 used for IoT sensor device communication. The 270 sensing elements and the corresponding electrical
224 majority of IoT sensor data incorporate real‐time 271 signal.
225 processing for industrial applications, healthcare, and
272 Fig.2 Sensing elements and corresponding electrical signals
226 scientific activities. The sensed data must 273 be There are different types of sensors which can range
227 processed to remove uncertainties for further data 274 from very simple to complex. The classification of
228 analysis, so as to develop knowledge and decision‐ 275 sensors can be based on their specifications, its
229 making. Thus, the data processing layer targets
230 different functions, such as data denoising, data
231 outlier detection, missing data imputation and data
232 aggregation. The data processing layer targets
233 different functions, such as data denoising, data
234 outlier detection, missing data imputation and data
235 aggregation. The data fusion layer is required to
236 handle various sensor data challenges generated by
237 several heterogeneous sensor devices. The data fusion
238 data aim to integrate true sensor data from
239 heterogeneous IoT sensor devices. The combined
240 data from different sources are then passed to the data
241 analysis layer for efficient knowledge generation and
242 decision-making. The primary data fusion involves
243 the direct fusion of data sensor data from different
244 sensor devices. It incorporates initial feature
245 extraction which is followed by data fusion. The
246 enhanced method involves feature extraction
247 276
followed by identity declaration and data fusion. This conversion method, type of material used, its sensing
277 physical phenomenon, properties that what 326 it A motion detector is a device used to sense all the
278 327
measures, and the application field. Fig. 3 depicts kinetic and physical movement in the environment.
279 328
various types of sensors in IoT which are explained An application for monitoring homes in the absence
280 below. 329 of homeowner can make use of motion sensors and
281 Fig.3 Different types of IoT Sensors 330 whenever motion is detected the photos or videos can
331 be uploaded on the server. Authors in used the motion
282 B. Types of Sensors 332 detection sensor for home security. The python script
333 is executed in the proposed approach to detect the
283 There are different types of sensors which can range
334 motion and this application takes continuous
284 from very simple to complex. The classification335 of snapshots till motion is detected and also captures
285 sensors can be based on their specifications, 336 its videos from the beginning of motion till the end of a
286 337
conversion method, type of material used, its sensing motion.
287 physical phenomenon, properties that what it
288 measures, and the application field. 338 5. Velocity sensors

289 1. Proximity Sensors 339 It is a sensor that calculates the rate of change in
340 constant position measurement and position values at
290 The position of any nearby object can easily 341 be known intervals. Velocity sensor may be linear or
291 detected with proximity sensor without any physical342 angular. A linear velocity sensor detects the speed of
292 contact. By emitting electromagnetic radiation such343 an object along a straight line whereas angular
293 as Fig 2 Different types of IOT sensors 344 velocity sensor detects how fast a device rotates. It
345 can be used in smart city applications for intelligent
294 infrared, it finds the presence of an object 346 by vehicle monitoring.
295 simply looking for any variation in the return
296 347
signal. There are different types of proximity sensors 6. Temperature sensors
297 like Inductive, Capacitive, Ultrasonic, Photoelectric,
348 Temperature sensors are helpful in detecting the
298 Magnetic and etc. targeting different applications.
349 physical changes in one’s body by measuring heat
299 This particular type of sensor is mostly used in
350 energy. Authors in used temperature sensors for the
300 applications demanding security and efficiency.
351 monitoring of environmental conditions of the
301 Various application areas of this type of sensors are
352 surroundings. The data collected is then sent to the
302 object detection, counting number of items,
353 cloud using Wi-Fi for the analysis. This is all done
303 measuring the amount of rotation, for positioning of
354 though android smartphone. A similar type of sensor
304 objects, material detection, measuring movement
355 is also used by the authors in for smart agriculture
305 direction, parking sensors and etc. Proximity sensors
356 and enable farmers to increase their overall yield and
306 have the best use in many types of industries.
357 product quality by getting real-time live data of their
307 b. Position Sensors 358 land.

308 The position sensor detects the presence of human359 or 7. Pressure Sensors
309 objects in a particular area by sensing their motion. It
310 can be used in home security to enable the owner to
360 Pressure sensors sense the amount of force and
311 track the doors and windows of rooms and appliances
361 convert it in signals. This type of sensors can be used
312 from anywhere. It let them know the open or closed
362 in health monitoring.
313 status at all times and can track the intruders in their
363 8. Chemical Sensors
314 absence. It can be used in health care monitoring for
315 monitoring the position of patients, nurses and 364 A chemical sensor is an analytical device used to
316 doctors in a hospital, in agriculture for detecting 365
the measure the chemical composition of the
317 position of cattle. 366 environment. Air quality monitoring can be done
318 3. Occupancy Sensors 367 using a wireless chemical sensor network by
368 monitoring chemical plumes in the environment.
319 Occupancy Sensor or sometimes called presence
320 sensor detects the presence of human or objects in369a 9. Humidity Sensors
321 particular area. It can be used for remote monitoring
322 through various parameters like temperature, 370 A humidity sensor measures air temperature as well
323 371
humidity light, and air. A similar application of this as moisture and signals the humidity in the
324 type of sensor is given by authors in. 372 environment. Authors in used humidity sensors for
373 smart agriculture and enable farmers to increase their
325 4. Motion Sensors 374 overall yield and product quality by getting real-time
375 live data of their land.
376 10. Water Quality Sensors 426 C. Smart sensors

377 Water quality sensors are used for Ion monitoring. 427 The embedded intelligence with basic sensing
378 Researchers in presented a design of a low-cost 428 mechanism is known as smart sensor. Five elements
379 system by measuring temperature, PH, turbidity, 429 for building smart sensor are sensing element,
380 conductivity, dissolved oxygen of the water and 430 memory, interface unit, signal processor unit and
381 monitored the water quality in the IoT scenario. 431 software. Different sensing elements detects signal
382 11. Infrared Sensors 432 for its processing, data interpretation, validity and
433 data logging. Smart sensors can accomplish access to
383 Infrared Sensors emits or detect infrared radiations434
in various facilities offered by smart cities as depicted in
384 order to sense some characteristics of certain objects.
435 Fig. 4. Several techniques implemented by sensors
385 They can also measure heat emission. This type436 of for collecting information through various sources are
386 sensors can be used for home automation for
387 monitoring and controlling home appliances like
388 turning on/OFF lights. It can also be used for smart
389 security, waste collection systems, smart parking.

390 12. Gyroscope Sensors

391 Gyroscope sensors detect any tilt or angular

392 movement in the object by measuring angular
393 velocity. It is widely used in 3D mouse games, for the
394 training of sportspersons, robotics, industrial
395 automation and many more.

396 13. Optical Sensors

397 Optical sensors are useful in detecting the

398 electromagnetic energies like light. Being passive to
399 all forms of electrical interfaces, these are widely
400 used in IoT applications like in digital cameras.
401 Optical sensors are good for IoT applications related
402 to energy, health, environment, oil refineries,
403 chemical, industries, aerospace and etc. 437 under continuous observation.
404 14. Chemical Sensors 438 Fig.4 Smart city benefits

405 Sensors which response by sensing any chemical 439 Sensors using silicon for their development as
406 reaction, chemical substance or a set of chemicals440
is possessed by piezo resistive, porous silicon and
407 441
known as chemical sensors. This type of sensors can MEMS. Sensors are fabricated via different
408 442
be used for detecting environmental events, building transduction techniques as shown in Table 1. These
409 health, agriculture conditions, and etc. 443 techniques provide an ease to fabricate, design,
444 operate and low cost in context of sensors
410 Having meticulously explored the diverse landscape 445 development sensors are capable in detection of
411 of conventional sensors employed in the Internet446 of electronic circuit, industries and environmental
412 Things (IoT), we arrive at a crucial juncture where447 applications.
413 the inherent limitations of these workhorses begin to
414 impede the full potential of data acquisition. Issues Sr.No. Transduction Technique
415 like cost, complexity, and the inability to sense 1. Micro gravimetric
416 specific parameters often hinder optimal 2. Electro Chemical
417 performance in dynamic IoT environments. 3. Electrolytic
418 Fortunately, the realm of sensor technology is 4. Optical
5. Resistive
419 constantly evolving, and on the horizon stand two
6. Capacitive
420 ground breaking advancements: smart sensors and
7. Impedance
421 virtual sensors. These ingenious offerings promise to
8. Piezoelectric Surface Acoustic Wave
422 revolutionize data collection by addressing the
423 shortcomings of traditional sensors and unlocking 448 TABLE I. Transduction techniques of sensors fabrication
424 exciting new possibilities for the ever-expanding
425 world of IoT.
449 The sensors are designed to provide healthy and 503 with each other, lose accuracy over time, or their use
450 504
protected life for humans, flora and fauna. Secondly, is eve technically not feasible, due to spatial or
451 505
for checking the superiority of air, irrigation soil, environmental conditions. On the other hand, virtual
452 506
lakes and rivers. Further, sensors reserve the natural sensors replace a subset of physical sensors with
453 resources for human sustainability. A systematic 507 virtual ones, allowing for the monitoring of
454 508
review is required on rainfall, volcano eruption, flush unreachable locations, reducing the sensors
455 floods, soil erosion, and many more. 509 deployment costs, providing a fallback solution, and
510 finally, improving the reliability of physical systems.
456 D. Virtual sensors
511 There are three main types of approaches to designing
457 Imagine an "invisible sensor" seamlessly gathering 512 a virtual sensor model:
458 vital data without physical presence. This isn't 513 a) mechanism-based
459 science fiction; it's the reality of virtual sensors,
460 revolutionizing the landscape of sensing in 514 the b) knowledge-based
461 Internet of Things (IoT). Virtual sensors leverage the
515 c) data-driven.
462 power of data and algorithms to estimate
463 environmental or system parameters, offering unique 516 What distinguishes these three types of approaches is
464 advantages and posing interesting challenges within 517 how the relation between the input parameters and the
465 the broader sensor ecosystem. Let's delve into this 518 output sensor value is defined. The relations can be
466 innovative technology and explore its impact on 519 the based on physical laws, expert knowledge, or a model
467 future of IoT. Virtual sensors have applications and 520 that recognizes data patterns. Specifically, the
468 multiple impacts in domains such as health-care, 521 mechanism-based virtual sensors are constructed
469 entertainment, fitness, sport, digital twins, and
522 based on the behavior of the operating mechanisms to
523 be measured, describing the essential physical
470 Industry 4.0. The first impact to mention is that they
524 correspondence between the input and output
471 extend the reliability and availability in applications,
525 quantities. The knowledge-based sensors use the
472 without adding extra hardware complexity and 526 knowledge of experts who explicitly define the
473 maintenance. They can be the basis for a fault 527 equations or rules between the input values and the
474 detection mechanism recognizing discrepancies 528 sensor output. The data-driven methods have become
475 between the sensor data and expected measurements. 529 the mainstream approach for training AI mechanisms
476 They provide an alternative when a physical sensor 530 based on historical data. Most data-driven virtual
477 cannot be placed in the preferred position, due531 to sensors apply a data fusion technique to provide
478 532
spatial conditions, such as a lack of space or a hostile precise measurements of one specific phenomenon or
479 environment, such as exposure to acids and extreme 533 an abstract representation of diverse sensor inputs.
480 temperatures. Virtual sensors can replace physical 534 This paper’s proposed virtual sensor methodology,
481 sensors and deliver a higher level of information 535 instead of aggregating the inputs of physical sensors,
482 based on multiple heterogeneous sensor signals.
536 replaces every single physical sensor with a virtual
537 one. They also use virtual sensors based on ML
483 Virtual sensors also have the impact that they
538 models to estimate the temperatures in indoor
484 overcome a number of weaknesses of physical 539 environments.
485 sensors, such as reducing signal noise and drifts.
486 Drifts constitute a well-known phenomenon 540
487 rendering a physical sensor inaccurate over time, due541 III. SENSOR DATA PROCESSING LEVELS FOR
488 to wear or pollution. The last impact to mention542 is VARIOUS IOT APPLICATION DOMAINS
489 that virtual sensors are extremely flexible and can be
490 redesigned as required, while physical sensors, once 543 Some IoT application domains, such as real-time
491 installed, often can only be repositioned 544 by applications, demand low latency, and high data
492 mechanical intervention. Sensor is a device which 545 transfer rates. Hence, sensor data processing can be
493 measures a physical quantity and transforms it into 546 performed at two different levels depending on the
494 sensor data that can be interpreted by an instrument 547 application requirements:1) near to the sensor data
495 or an observer. Next, they distinguish the physical 548 origin (node level) or 2) at the cloud level. Further,
496 sensors, which measure physical phenomena directly 549 node-level sensor data processing is categorized into
497 from the virtual sensors that are software-based 550 a) Edge computing, and b) Fog computing.
498 models. The virtual sensors produce sensor data 551 by A. Data Processing at/Near the Sensor Node:
499 fusing data received synchronously or
500 asynchronously from physical or other virtual 552 Edge Computing: The edge computing permits the
501 sensors. Comparing physical to virtual sensors shows 553 data to be handled close to its cause (the sensor
502 554
that physical sensors may be prone to noise, interfere devices/gadgets). The information is moved from
555 sensor gadgets to a neighbourhood edge computing 568 (sensor devices) to complete a considerable measure
556 569
framework, which procedures, stores the information, of calculation, storage, and communication locally
557 and processes the data locally. Additionally, 570 the before being transmitted over the web. In 2014, the
558 framework could assemble the handled information 571 need to expand distributed computing with fog
559 572
and send it to the Cloud at regular intervals of time. computing rose, and the idea was given by Cisco. Fog
560 The beneficial thing about edge processing is that 573 computing includes carrying knowledge to the local
561 only the significant data is sent over the network.574 network of computing devices and processes the data
562 575
This requires less data transmission from the sensor in a fog node. Fig.5 depicts the generic sensor data
563 system and spares the sensor batteries. Likewise, 576
the processing at various. As indicated by Cisco, fog
564 information can be processed faster when done close577 computing is a standard that characterizes how edge
565 to the sensor gadget. 578 computing should work, and it empowers the activity
579 of process,
566 Fog Computing: Fog Computing, otherwise called
567 fogging, is engineering that utilizes edge gadgets

580 Fig.5 Sensor Data processing levels node (Edge Flash/fog), cloud level

581 storage, and systems administration between end 594 cloud, is significant for data aggregation,
582 595
gadgets/devices and distributed computing systems. subsequently inferring intelligent decisions from the
583 Fog computing underpins the IoT idea, in which596 a information. For example, an agribusiness
584 large portion of the devices utilized by people 597 organization would have the option to look at soil
585 consistently will be associated with one another.598 moisture from the sensors located at different places
586 599
Models incorporate telephones, wearable well-being after planting similar seeds. Without the Cloud,
587 600
checking gadgets, associated vehicles, and wearable looking at information across more extensive zones is
588 601
devices (e.g., Google Glass). IEEE received the Fog substantially more troublesome. Utilizing the cloud
589 Computing norms proposed by Open Fog Consortium 602 computing features for IoT takes into account high
590 for standardization. 603 versatility. When a massive number of sensors is
604 utilized, putting a high computational burden on
591 B. Sensor Data Processing at the Cloud: Numerous605 every sensor would be very costly. Instead,
592 IoT frameworks utilize vast sensors to gather
593 information and make smart decisions. Utilizing the
606 information can be passed to the Cloud from every 659 management tasks improves the designs of
607 sensor and prepared there in total. For quite a bit660
of interconnected IoT systems. Innovations (AI, bots,
608 IoT, sensors and gadgets gather information and 661 mechanization, AR/VR) make intelligent decisions
609 perform activities, yet the handling/ordering 662 through combined IoT knowledge thereby expanding
610 663
/investigation ordinarily occurs in the Cloud. Sensor human capabilities and increasing abilities of
611 data is continuously streamed to the Cloud. IoT 664 machines/things to better handle and manage the
612 665
applications that do not concern much on the latency edge/fog computing areas. A smart sensor is an AI
613 and bandwidth do prefer cloud computing scenario. 666 thing of what is known as a sense in science. With a
614 AI for sensors is getting simpler than any time in 667
the smart sensor, a machine monitors the environment
615 recent past equipment costs are decreasing, and 668 and data can be gathered. A sensor quantifies a
616 sensors are getting less expensive, making IoT 669 physical amount and converts it into a sign. Sensors
617 gadgets broadly accessible for an assortment 670 of interpret estimations from this present reality into
618 utilizations running for smarter decisions. 671 information for the computerized area. A practically
672 unbounded variety of boundaries can be estimated.
619 IV. APPLICATIONS OF IOT 673 Examples include area, uprooting, development,
674 sound
620 Imagine a world where everyday objects, from your
621 coffee pot to your city's traffic lights, are 675 Fig 4 Sensor IoT processing level Nodes
622 interconnected, collecting and sharing data. This
623 676
network of intelligent devices, known as the Internet recurrence, temperature, pressure, dampness,
624 of Things (IoT), is revolutionizing various aspects677
of electrical voltage level, camera pictures, shading,
625 678
our lives. IoT sensors in agriculture can monitor soil synthetic creation, and so on. The objective is to
626 679
moisture and nutrient levels, optimizing crop yields. identify functions or changes in the environment.
627 IoT can optimize energy consumption in buildings, 680 Different sensors are commonly used in different
628 leading to cost savings and environmental benefits. 681 gadgets. Advances in chip designs and
629 From enhancing our personal comfort and 682 implementation makes it possible to develop small,
630 convenience to optimizing large-scale operations, 683
the cost-effective, and energy-efficient sensors. The
631 applications of IoT span across industries and hold 684 quantity of sensors around us is expanding quickly
632 685
immense potential for shaping a smarter future. Let's and as IoT grows further in coming years, the number
633 delve into the diverse ways this technology 686 is and types of sensors that will be deployed in the IoT
634 transforming our world. 687 environment will continue to rise. A smart sensor
635 A. Artificial Intelligence (AI) 688 alludes to the incredible significance of completing
689 estimations to make concrete and verifiable data
636 690
From simple data collection to intelligent analysis, accessible. Measurement (estimation) of data through
637 the role of sensors in the IoT has evolved 691 smart sensors is very much needed. Estimating gives
638 692
significantly. With the advent of AI, this evolution an understanding of things that work out in the right
639 takes a quantum leap. 693 way.

640 This review explores the historical integration of 694

AI 1. Current Applications of AI in IoT
641 and sensors in the IoT, highlighting key milestones
642 and advancements. We then delve into the future 695 Presently, AI is actively contributing to diverse
643 potential of this powerful duo, analysing its impact696
on applications within IoT ecosystems. Operational
644 shaping the next generation of intelligent and 697 efficiency receives a boost as AI algorithms
645 autonomous systems. AI, encompassing Machine 698 streamline processes and predict maintenance needs
646 Learning, Deep Learning, and Natural Language 699 in industrial setups. In smart buildings, AI optimizes
647 700
Processing, is increasingly integrated into IoT, paving energy consumption, reducing environmental impact.
648 the way for Artificial Intelligence of Things (AIoT)701 Risk management benefits from AI's analytical
649 applications. Currently, AI in IoT boosts efficiency702 capabilities, ensuring worker safety through the
650 (e.g., Google's data Centre cooling), improves risk 703 analysis of data from wearable devices. Scalability is
651 management (e.g., Fujitsu's worker safety analysis),704 addressed as AI tackles the challenges posed by the
652 enables new products (e.g., Rolls Royce's engine 705 sheer volume of data in expansive IoT deployments.
653 maintenance), and enhances scalability. The future 706 2. Security and Autonomy in IoT
654 holds promise for AI-driven edge data reduction,
655 smart object personality design, and improved 707 The inseparable connection between AI and IoT
656 security through self-governing activities. The mix708
of extends to security measures and autonomy. AI's role
657 AI-based techniques at various IoT compositional 709 in enhancing security involves anomaly detection and
658 layers to deal with IoT information for self- 710 threat identification, mitigating vulnerabilities in
711 763
interconnected systems. Furthermore, AI enables configuration tools for smart sensing systems,
712 autonomous decision-making, fostering 764
self- emerges as key solutions to enhance integration
713 governance within IoT environments. These 765 efficiency. Within the domain of latency and
714 766
advancements not only improve system robustness scalability, the demand for low-latency responses in
715 767
but also contribute to the realization of fully many IoT applications necessitates a delicate balance
716 autonomous IoT operations. 768 between responsiveness and efficiency. While data
769 storage optimization and architectural considerations
717 3. Smart Sensors in IoT 770 contribute to distributed intelligence, challenges
718 Smart sensors emerge as the linchpin of IoT 771 persist in configuring smart sensing systems within
719 772
functionality, converting real-world physical data into the AIoT framework. Addressing these intricacies
720 digital information. Recent strides in sensor 773 requires focused efforts on system setup and
721 technology encompass miniaturization, energy 774 configuration to ensure optimal performance. The
722 efficiency, and enhanced sensitivity. These sensors 775 challenge of quantum resilience unfolds as the need
723 play a pivotal role in diverse applications, from 776 to ensure quality services from a multitude of
724 monitoring environmental conditions to enabling 777 connected devices becomes increasingly complex.
725 precise data acquisition for complex IoT analytics. 778 Managing the intricacies associated with trillions of
779 connected sensors necessitates strategic approaches to
726 4. Future Research Directions: 780 enhance responsiveness, scalability, processes, and
781 overall efficiency. The integration of quantum
727 Looking ahead, the future integration of AI and IoT
782 technology emerges as a promising solution to
728 holds promising avenues. Designing data reduction
783 various aspects, including network latency,
729 approaches for embedded edge nodes is on the
784 interoperability, real-time analytics, predictive
730 horizon, addressing the need for efficient processing
785 analytics, increased storage and data memory, secure
731 and storage. Additionally, the concept of infusing
786 cloud computing, and the demands emerging 5G
732 human-like personalities into smart objects envisions
787 telecommunications infrastructure. Investing in
733 a more intuitive and user-centric IoT experience. As
788 research and development for the seamless
734 these ideas evolve, ethical considerations will play a
789 integration of quantum technologies into AI-based
735 crucial role in shaping the future landscape of AI in
790 sensor systems is crucial, promising solutions to
736 IoT.Future research should delve into developing
791 challenges related to network latency, real-time
737 energy-efficient AI algorithms tailored for edge
792 analytics, and scalability. Additionally, fostering
738 computing in IoT. The impact of quantum computing
793 interdisciplinary collaboration and knowledge
739 on addressing complex IoT data challenges warrants
794 exchange stands out as a holistic approach to
740 exploration. collaborations can bridge knowledge
795 addressing the multifaceted challenges posed by
741 gaps and foster innovative solutions, ensuring that
796 AIoT, unlocking the full potential of AI based sensor
742 AIoT continues to evolve to meet the dynamic needs
797 systems across diverse application domains.
743 of our interconnected world.
798 B. Industry 4.0: The Transformative Role of Sensors
744 5. Challenges and Solutions in the Integration of AI
799 in Smart Manufacturing
745 with IoT
800 The introduction contextualizes the evolution of
746 The convergence of Artificial Intelligence (AI) with
801 Industry 4.0, emphasizing its transformative influence
747 the Internet of Things (IoT) presents a myriad of
802 on manufacturing. It introduces sensors as pivotal
748 challenges, each requiring thoughtful consideration
803 components, essential for realizing the vision of
749 and innovative solutions. In the realm of AIoT
804 Industry 4.0 where interconnected machines
750 adoption and interoperability, the proliferation of AI-
805 collaborate intelligently in real-time.
751 based sensor applications has resulted in a landscape
752 of heterogeneous solutions, complicating efforts 806 1. Core Principles of Industry 4.0:
753 towardsseamless integration. Diverse
754 implementations and features 807
introduce This section delves into the fundamental principles
755 interoperability challenges, obstructing 808
the that define Industry 4.0, such as automation,
756 establishment of standardized practices. Moreover, 809 connectivity, data-driven decision-making, increased
757 the complexity of APIs provided by Edge/Fog 810 efficiency, and adaptability. It explores how sensors
758 811
computations adds to usability concerns, necessitating act as catalysts in implementing these principles,
759 812
solutions that prioritize interoperable frameworks and facilitating seamless communication between
760 standardized protocols. Simplifying APIs through 813 devices, enabling automation, and driving informed
761 improved usability design and standardized 814 decision-making.
762 interfaces, along with the development of advanced
815 2. Sensor Technologies in Industry 4.0: 865 The Internet of Things (IoT) is a communication
866 paradigm that envisions a future in which everyday
816 A detailed exploration of various sensor technologies
867 objects are equipped with microcontrollers,
817 is undertaken, showcasing their adaptability and 868 transceivers, and protocol stacks to communicate
818 versatility. The review emphasizes the role of sensors
869 with each other and with users, becoming part of the
819 in real-time data gathering, remote surveillance, and
870 Internet. The IoT concept aims to make the Internet
820 preventive maintenance, revolutionizing equipment 871 even more immersive and pervasive, enabling access
821 performance across diverse industries. Additionally,
872 to a wide variety of devices to provide new services
822 it discusses emerging sensor technologies such 873 as to citizens, companies, and public administrations. It
823 LiDAR, RFID, and advanced imaging techniques, 874 finds application in various domains such as home
824 expanding the scope of applications in Industry 4.0.875 automation, industrial automation, medical aids,
825 3. Data-Driven Nature of Industry 4.0: 876 mobile healthcare, elderly assistance, intelligent
877 energy management, smart grids, automotive, traffic
826 This section highlights the data-driven nature 878 of management, and more.
827 Industry 4.0 and how sensors contribute to this ethos.
828 879
It discusses the role of sensors in providing raw data, However, due to the extreme variety of devices, link
829 aiding in the identification of inefficiencies, and880 layer technologies, and services involved in IoT
830 enhancing machine performance. The review 881 systems, building a general architecture for the IoT is
831 emphasizes the significance of smart sensing 882 a complex task. Specifically, applying the IoT
832 technologies in collecting and analysing large 883 paradigm to an urban context is of particular interest,
833 884
volumes of data rapidly and accurately, facilitating aligning with the Smart City concept to use ICT
834 predictive analytics and machine learning 885 solutions for managing public affairs. The Smart City
835 applications. 886 concept aims to utilize advanced communication
887 technologies to provide added-value services and
836 4. Challenges and Solutions: 888 improve the quality of services for citizens while
837 The paper acknowledges the challenges associated 889 reducing operational costs for public administrations.
838 with the integration of sensors in Industry 4.0, 890 This paper provides a comprehensive survey of the
839 including cybersecurity threats, data privacy 891 enabling technologies, protocols, architecture for an
840 concerns, interoperability issues, and the high cost892
of urban IoT, as well as technical solutions and best-
841 sensor deployment. To address these challenges,893 it practice guidelines adopted in the Padova Smart City
842 proposes robust cybersecurity measures, encryption 894 project, a proof-of-concept deployment of an IoT
843 895
protocols, industry-wide standards, and collaborative island in the city of Padova, Italy, in collaboration
844 efforts to ensure secure and standardized 896 with the city municipality.
845 communication between sensor-equipped devices.
846 Additionally, it explores cost-effective sensor 897 1. SMART CITY CONCEPT AND SERVICES
847 deployment strategies and emphasizes the importance 898 The services commonly associated with the Smart
848 of fostering a regulatory framework to address 899 City vision include Smart Governance, Smart
849 privacy concerns. 900 Mobility, Smart Utilities, Smart Buildings, and Smart
850 5. Opportunities and Future Directions: 901 Environment. However, political, technical, and
851 The paper explores opportunities presented by 902 the financial barriers have hindered the Smart City
852 integration of sensors in Industry 4.0, focusing 903
on market from taking off. The lack of interoperability
853 predictive maintenance, asset monitoring, enhanced904 of heterogeneous technologies used in urban
854 product design, and the potential for creating new905 developments, as well as the absence of a clear
855 business models. It discusses the role of sensors906
in business model, contribute to the obstacles faced in
856 enabling sustainability practices through energy 907 the successful deployment of Smart City solutions.
857 efficiency and waste reduction. The conclusion 908 [14]
858 contemplates future directions, potential
909 Despite the barriers, multiple services can be enabled
859 advancements, and the evolving role of sensors910 in by an urban IoT paradigm, including structural health
860 smart manufacturing, encouraging further research911
to of buildings, waste management, air quality, noise
861 harness the full potential of sensors for a more 912 monitoring, traffic congestion, city energy
862 efficient, intelligent, and sustainable manufacturing
913 consumption, smart parking, smart lighting, and
863 future. 914 automation and salubrity of public buildings. Each of
864 C. Smart cities and homes 915 these services presents unique capabilities,
916 challenges, and possibilities for improving the quality
917 of life for citizens in urban environments. The
918 practical realization of these services requires944 Representational State Transfer (ReST) paradigm.
919 945
deploying sensors, integrating device powering, and This approach facilitates the adoption and use of IoT
920 946
establishing communication infrastructure for data by end users and service developers, enabling them to
921 collection and processing.[14] 947 reuse knowledge gained from traditional web
948 technologies. The protocol architecture for urban IoT
922 The integration of an urban IoT can enable a wide 949 systems includes both unconstrained and constrained
923 range of services, improve the quality and efficiency
950 protocol stacks, facilitating easy access and
924 of public services, and create potential economic 951 interoperability of IoT nodes with the Internet.
925 advantages for city administrations. By leveraging the
926 IoT paradigm, Smart City initiatives can thrive and952 technologies with existing communication
927 help cities become more sustainable, efficient, and953 infrastructures to support progressive evolution and
928 citizen-friendly.[14] 954 interconnection of devices. It emphasizes easy
955 accessibility of data for authorities and citizens, and
929 2. URBAN IOT ARCHITECTURE 956 utilizes web service approaches and IETF standards
930 Urban IoT services are typically based on a
931 centralized architecture with a dense and diverse set
932 of peripheral devices generating data across the urban
933 area. This data is then transmitted through
934 communication technologies to a control centre for
935 storage and processing. The main challenge for urban
936 IoT infrastructure is its ability to integrate different
937 technologies with existing communication
938 infrastructures to support progressive evolution and
939 interconnection of devices, along with easy
940 accessibility of the data for authorities and
941 citizens.The web service approach for IoT services is
942 based on IETF standards, which embrace a web
943 service architecture designed in accordance with the

957
958 In terms of data format, the urban IoT paradigm sets 983 Constrained and unconstrained link layer
959 specific requirements for data accessibility, leading984
to technologies, such as Ethernet, WIFI, and IEEE
960 the development of new technologies such as EXI 985 802.15.4, facilitate the realization of an urban IoT
961 (Efficient XML Interchange), which allows even very 986 system across a wide geographical area. Mobile
962 constrained devices to support and generate messages 987 devices, such as smartphones and tablets, can also
963 using an open data format compatible with XML. 988 play an important role in providing access to the IoT
964 Integration of multiple XML/EXI data sources into989 an and interacting with IoT objects.
965 IoT system is achieved through databases maintained
966 by high-level applications. Additionally, the
967 application and transport layers of urban IoT systems
968 are supported by CoAP (Constrained Application
969 Protocol) as an alternative to traditional HTTP over
970 TCP for constrained IoT devices.

971 At the network layer, IPv6 is the leading addressing

972 technology supported by IoT networks, making it
973 possible to assign a unique address to every node in
974 the IoT network. Additionally, 6LoWPAN is an
975 established compression format for IPv6 and UDP
976 headers over low-power constrained networks. In
977 terms of devices, an urban IoT system consists of
978 backend servers located in the control centre,
979 gateways that interconnect end devices to the main
980 communication infrastructure, and IoT peripheral
981 nodes responsible for generating data to be delivered
982 to the control centre.
990 In summary, the urban IoT architecture is based 1031
on a constrained link layer technology, with routing
991 centralized infrastructure, integrating different 1032 provided by the RPL protocol. The gateway
1033 interfaces with the sensor cloud and connects it to
992 to enable interoperability of IoT nodes with 1034 the backend servers using traditional WAN technologies.
993 Internet. Additionally, it leverages new technologies
1035 The HTTP-CoAP proxy enables communication with
994 such as EXI and CoAP to support constrained 1036 IoT CoAP devices, allowing monitoring and traffic
995 devices and address unique requirements for 1037 data shaping. The database server collects and stores data
996 accessibility and data exchange. 1038 from the sensors and can be accessed through web
997 1039
3.AN EXPERIMENTAL STUDY: PADOVA SMART programming technologies. Public lighting operators
998 CITY 1040 are equipped with mobile devices to monitor and
1041 control streetlamps. Figs. 6 and 7 report an example
999 The framework discussed in this paper has already 1042 of the type of data that can be collected with the
1000 been successfully applied to a number of different1043
use Padova Smart City system.
1001 cases in the context of IoT systems.” For example,
1002 the University of Padova has deployed a wireless
1003 sensor network testbed with over 300 nodes
1004 according to these guidelines, which has been used
1005 for proof-of-concept demonstrations of smart grid and
1006 health care services. The "Padova Smart City"
1007 project, implemented in the city of Padova, is a
1008 practical implementation of an urban IoT. It is a
1009 system for collecting environmental data and
1010 monitoring public street lighting using wireless nodes
1011 1044
with various sensors. These nodes are placed on street
1012 light poles and connected to the internet through1045 a Fig.6. Example of data collected by Padova Smart City: (a)
1013 gateway unit. The system collects parameters such as
1046 temperature and (b) humidity.
1014 CO level, air temperature and humidity, vibrations,
1015 noise, and measures the light intensity at each post to
1016 check the correct operation of the public lighting
1017 system.

1018 A conceptual sketch of the

1019 Padova Smart City system architecture is given in
1020 Fig. 5. The components of the Padova Smart City
1021 system include street lights, IoT nodes, constrained
1047
1022 link layer technologies, a WSN gateway, an HTTP-
1048 Fig.7. Example of data collected by Padova Smart City: (a) light
1023 CoAP proxy, and a database server. Each streetlight1049 and (b) benzene.
1024 is associated with an IoT node that collects data and
1025 provides context information. The IoT nodes are 1050 The Padova Smart City system collects various types
1026 1051 of data, including temperature, humidity, light
1027 Fig.5 system architecture of Padova city 1052 intensity, and benzene levels. The data is plotted over
1053 a period of 7 days, showing patterns and variations.
1028 powered by small batteries and equipped with sensors
1054 For example, light intensity follows a regular pattern
1029 for temperature, humidity, and air quality. They form
1055 of day and night, while benzene levels decrease at

1030 1056
a 6LoWPAN multi hop cloud using IEEE 802.15.4 nighttime due to lighter traffic. The system provides
1057 valuable insights into environmental conditions 1107
and virtual town halls to involve residents in decision-
1058 1108
can optimize maintenance plans based on real-time making processes.
1059 data.
1109 f. Resilience and Security: Smart cities will invest in
1060 Overall, the Padova Smart City project demonstrates 1110 technologies to enhance resilience against natural
1061 the successful implementation of an urban IoT system1111 disasters, cyber-attacks, and other threats. This
1062 for collecting environmental data and monitoring 1112 includes the use of IoT sensors for early warning
1063 1113
street lighting. It showcases the use of wireless nodes, systems, blockchain for secure data management, and
1064 constrained link layer technologies, gateway 1114 AI for threat detection. Ensuring the security of smart
1065 interfaces, and data storage solutions. The system1115
has city infrastructure is crucial to prevent unauthorized
1066 the potential to be expanded to include other types of
1116 access and manipulation.
1067 IoT nodes and can be controlled through an open,
1068 standard, and technology-independent interface. 1117 g. Health and Well-being: Smart cities will prioritize
1118 initiatives that promote health and well-being, such as
1069 4. FUTURE DIRECTIONS 1119 smart healthcare systems, green spaces, and active
1120 lifestyle programs. This includes the use of wearables
1070 Future directions for smart cities are likely to focus
1121 and health tracking devices to monitor and improve
1071 on further integration of technology to enhance
1122 public health.
1072 sustainability, efficiency, and quality of life for
1073 residents. Here are some key areas that are expected
1123 Overall, the future of smart cities will be
1074 to drive the evolution of smart cities: 1124 characterized by greater integration of technology,
1125 data-driven decision-making, and a focus on
1075 a. Sustainable Infrastructure: Smart cities will
1126 sustainability and citizen-centric services.
1076 increasingly prioritize sustainable practices, such as
1077 renewable energy sources, efficient waste 1127 D. Smart Farming
1078 management, and green building designs. This
1079 includes the use of IoT sensors and data analytics1128
to The challenge of food production in the 21st century
1080 optimize resource usage and reduce environmental 1129 is a growing concern due to the increasing global
1081 impact. 1130 population, estimated to reach between 9.4 and 10.1
1131 billion people by 2050. This will lead to a higher
1082 b. mobility and Transportation: Future smart cities
1132 demand for food production areas, while
1083 will prioritize mobility solutions that reduce 1133 environmental changes caused by human activity may
1084 congestion, improve public transportation, 1134 and limit the development of new crops. Smart farming, a
1085 promote alternative modes of transportation such as
1135 new farm management concept, utilizes techniques
1086 electric vehicles, bike-sharing, and pedestrian-
1136 and technologies to overcome these challenges by
1087 friendly infrastructure. This may also include1137
the employing sensors to collect data and communication
1088 integration of autonomous vehicles and smart traffic
1138 networks to manage and analyse the data. This
1089 management systems. 1139 system, known as the Internet of Things (IoT), helps
1140 in boosting productivity and reducing waste. Recent
1090 c. Digital Connectivity: A key focus will be on
1141 technological developments have made it easier to
1091 expanding digital connectivity to ensure reliable,
1142 adopt smart farming with IoT, and the world’s largest
1092 high-speed internet access for all residents. This
1143 agricultural producers are promoting the usage of IoT
1093 includes the deployment of 5G networks, public Wi-
1144 in smart farming through incentive programs and
1094 Fi hotspots, and smart city platforms that enable
1145 public policies.
1095 seamless digital interactions and services.
1146 Several reviews have been published on IoT solutions
1096 d. Data-driven Governance: Smart cities will
1147 for smart agriculture in recent years, focusing on
1097 increasingly rely on data analytics and AI to improve
1148 topics such as network technologies, communication
1098 decision-making processes and enhance public
1149 protocols, data generation and analysis, IoT
1099 services. This includes predictive analytics for crime
1150 architectures, and applications. These reviews have
1100 prevention, real-time monitoring of urban
1151 identified an increasing use of modern techniques and
1101 infrastructure, and personalized services based on
1152 technologies for processing the collected data, such as
1102 citizen data.
1153 big data, machine learning, computer vision, and
1103 e. Citizen Engagement: Future smart cities 1154 will blockchain. This study aims to identify how IoT is
1104 emphasize citizen engagement and participatory 1155 used in smart farming by conducting a
1105 1156
governance through digital platforms. This includes comprehensive review of the state-of-the-art IoT
1106 tools for public feedback, community forums, 1157
and adoption in smart agriculture. The study surveys
1158 literature until the year 2020 and evaluates the
1159 acceptance of new IT technologies, such as big data,1193 processing layers using network protocols. The
1160 1194
computer vision, artificial intelligence, blockchain, processing layer comprises data storage,
1161 and fuzzy logic, in the smart agriculture field. 1195
The visualization, and processing resources, including big
1162 review focuses on the agriculture economic sector, 1196 data, artificial intelligence, and machine learning. The
1163 including indoor and outdoor agriculture, 1197 but application layer provides management information
1164 excludes livestock farming due to its specificities.1198 to farmers, enabling them to manage the entire
1199 production process in the plantations.
1165 Smart farming involves the application of
1166 technological resources to agricultural production1200 3. Methods
1167 techniques to minimize waste and boost productivity.
1168 1201
This includes the use of various technologies, such as The study used the Preferred Reporting Items for
1169 temperature and humidity sensors, unmanned flying 1202 Systematic Reviews (PRISMA) methodology to
1170 equipment, and global positioning systems,1203 to conduct a systematic review of literature related to
1171 monitor plantations, soil management, irrigation,1204
and the application of IoT in smart farming. A search was
1172 pest control. The integration of technological 1205 conducted in the Scopus database in 2019 and 2020,
1173 resources into the agricultural production process1206
is resulting in 463 articles. After screening and
1174 1207
expected to optimize food production and reduce the eligibility phases, 159 articles were selected as a
1175 use of pesticides and water consumption. 1208 sample for the study.
1209 Fig.8. PRISMA flowchart of the systematic review on state-of-the-
1176 1. Internet of Things (IoT) 1210 art IoT solutions in smart farming

1177 IoT is a network of interconnected intelligent devices

1211 4. Discussion
1178 capable of communicating with each other 1212 and Based on the results obtained in the analysis of the
1179 generating relevant data about the environment. 1213
The articles considered for this study, it was possible to
1180 adoption of IoT has increased in recent years due to
1214 observe a growth trend in the number of publications
1181 advancements in hardware, connectivity, cloud 1215 related to IoT and smart farming since 2011, with
1182 computing, artificial intelligence, and big data. 1216
The special emphasis from 2016 onwards, as shown in
1183 architecture of IoT systems consist of four layers:
1217 Figure 3.The analysis of the selected articles showed
1184 devices, network, services, and application, enabling
1218 a significant increase in the number of publications
1185 data sharing and active interactions between devices.
1219 related to IoT and smart farming since 2011, with a
1186 2. Intelligent Agriculture
1220 notable increase from 2016 onwards. The growing
1221 number of published papers highlights the increasing
1222 discussion and relevance of IoT applied to smart
1223 farming.
1224 This systematic review provides valuable insights
1225 into the state-of-the-art IoT adoption in smart
1226 agriculture, shedding light on the growing use of
1227 modern technologies and techniques to address the
1228 challenges in food production. It contributes to the
1229 existing literature by identifying the increasing
1230 adoption of new IT technologies in the smart
1231 agriculture field, such as big data, computer vision,
1232 artificial intelligence, blockchain, and fuzzy logic.
1233 This comprehensive review sets the stage for future
1234 research and advancements in the application of IoT
1187 The IoT architecture for intelligent agriculture
1235 in smart farming.
1188 includes perception, transport, processing, and
1189 1236
application layers. The perception layer collects data Fig.9. Classification of reviewed papers according to the year of
1237 application.
1190 from physical devices like sensor nodes and
1191 communication gateways, while the transport layer 1238 Environments for indoor scenario are protected from
1192 facilitates communication between the perception1239
and climatic impacts, such as solar radiation, rain and
1240 wind. Figure 4. Typical scenarios in agriculture1294
can collecting data from agricultural equipment such as
1241 be divided in indoor and outdoor. Examples1295 of tractors and harvesters to optimize maintenance and
1242 Fig.10.Ways of agriculture depending upon location 1296 performance.
1243 environments for indoor scenarios include 1297 Additionally, the reviewed papers mention the
1244 greenhouse, hydroponics, crop beds, pots, etc.1298 In challenges presented by agricultural environments,
1245 1299
contrast, environments for outdoor scenario are more such as communication obstacles due to distance,
1246 susceptible to climatic impacts. Examples 1300 of lack of communication in croplands, impact of
1247 1301
environments for outdoor scenario are arable lands, or vegetation on signal propagation, and the influence of
1248 Typical scenarios in agriculture can be divided 1302
in climatic elements on the planting and sensor nodes.
1249 indoor and outdoor. Indoor agriculture includes 1303 To address these challenges, 96% of the reviewed
1250 environments such as greenhouse, hydroponics 1304 and papers used commercial electronic sensors to collect
1251 1305
crop beds. Outdoor agriculture includes environments real-time data about various agricultural parameters,
1252 such as orchards and arable lands. IoT solutions1306
that such as climatic data, substrate information,
1253 1307
may be applied to multiple environments are referred luminosity, and images. Some papers (4%) focused
1254 to as” Generic.” 1308 on developing custom-made sensors for monitoring
1309 specific agricultural aspects, such as soil nutrients and
1255 5. Application:
1310 leaf evapotranspiration.
1256 Chemical control (e.g., pesticides and fertilizers).
1257 Crop monitoring. 1311 Different types of sensors were used in IoT solutions
1258 a) Disease prevention. 1312 for smart agriculture to collect data related to crop
1259 b) Irrigation control. 1313 monitoring, substrate monitoring, and environment
1260 c) Soil management. 1314 monitoring. For instance, electronic sensors were
1261 d)Supply chain traceability 1315 used for environmental data collection, such as
1262 e) Vehicles and machinery control. 1316 temperature, humidity, and luminosity, while cameras
1263 The most common applications of IoT solutions1317 for and multispectral sensors were used for crop
1264 smart farming are crop monitoring, disease 1318 monitoring to collect images of crops.
1265 prevention, irrigation control, soil management,
1266 supply chain traceability, and vehicles and machinery1319 The choice of hardware used in IoT project
1267 control. These solutions have been developed1320 for development is essential as it impacts costs and
1268 multiple agricultural environments, such as arable 1321 technologies available. Single Board Computers
1269 lands, orchards, and greenhouses. The widespread1322 use (SBCs), such as Arduino, Raspberry Pi, and ESP
1270 of these applications in agriculture is due to1323 the boards, were commonly mentioned hardware in the
1271 relevance of crop monitoring for farmers. 1324 IoT reviewed papers. These devices are affordable and
1272 solutions developed for monitoring crops are focused 1325 versatile, enabling the development of custom-made
1273 on collecting environmental data such as temperature,1326 IoT devices and the capability to work as gateways or
1274 humidity, and luminosity. The data collected is used 1327 sensor nodes in IoT solutions.
1275 to determine the vigour of crops and control 1328 Network protocols used in the reviewed papers
1276 environmental conditions. Similarly, IoT solutions1329 for included technologies like Wi-Fi, ZigBee, Bluetooth,
1277 irrigation control aim to optimize water resource1330 use CAN, Ethernet, LoRaWAN, GPRS, and 3G, among
1278 in agriculture, either through simple measurements 1331 others, to enable communication between IoT
1279 with sensors or more sophisticated approaches 1332 that solution devices. Table 5 shows the variety of
1280 combine humidity data with weather datasets. 1333 network protocols used in different agricultural
1281 Disease prevention solutions collect various 1334 environments to create short or long-range networks.
1282 environmental and plantation data, such as images of
1335 Additionally, it was noted that network topology is an
1283 plants, sounds, and temperature, to identify 1336 and essential aspect of an IoT solution, impacting the
1284 prevent diseases on plantations. 1337 distance between sensor nodes and the destination,
1285 Moreover, IoT solutions for chemical control aim 1338
to as well as the number of sensor nodes in the WSN.
1286 1339
optimize the application of fertilizers and pesticides The topology adopted by 61% of the reviewed papers
1287 on plantations based on crop data, such as nitrogen 1340 included star, tree (or cluster), and mesh network
1288 concentration in crops and aerial images that 1341 help topologies.
1289 determine specific regions needing fertilizer. For1342
soil Cloud-based platforms such as Thing Speak,
1290 management, IoT solutions are used to measure1343 soil FIWARE, Ubidots, Smart Farm Net, AWS IoT, and
1291 moisture, identify water consumption patterns, 1344 and Thinger.io were commonly used in the reviewed
1292 determine nutrient levels in the soil. Additionally,1345
IoT papers for processing data. These platforms offer
1293 solutions for vehicles and machinery control focus on
1346 1399
capabilities for data storage, processing, visualization, in combination with Single Board Computers (SBC)
1347 and action control. Furthermore, these platforms 1400 or Unmanned Aerial Vehicles (UAVs), such as
1348 provide scalability for IoT solutions by relying1401 on Arduino, Raspberry Pi, and drones. Different types of
1349 cloud computing to process and store large amounts 1402 network connections are used for IoT device
1350 of data obtained from sensors. The use of different 1403 communication. Wired networks like CAN and
1351 technologies like artificial intelligence, machine 1404 Ethernet are commonly used for indoor agriculture,
1352 learning, and big data supports data processing in1405
IoT such as in greenhouses, due to their resistance to
1353 solutions. Crop monitoring was the most common
1406 climate conditions and shorter distances between
1354 application for IoT solutions using data processing
1407 sensor nodes. Wireless connections, including Wi-Fi,
1355 technologies, and these technologies were also
1356 employed to support chemical control, disease
1408 are employed in both indoor and outdoor agriculture,
1357 prevention, and soil management applications in1409 IoT although power consumption and signal range
1358 solutions. 1410 limitations may restrict the use of Wi-Fi in larger or
1411 power-restricted projects. To overcome these
1359 the integration of IoT solutions into smart farming
1412 limitations, energy-efficient protocols like ZigBee,
1360 provides a wide range of applications to aid farmers
1413 BLE, and LoRa are used in wireless networks.
1361 in crop monitoring, disease prevention, irrigation
1362 1414
control, soil management, supply chain traceability, While the focus of this review was on IoT solutions
1363 1415
and vehicles and machinery control across various applied to agriculture, it is worth noting that IoT can
1364 agricultural environments. Through the use1416 of also be applied to other farming-related activities like
1365 commercial electronic sensors, single board 1417 livestock management. Power supply considerations
1366 1418
computers, and cloud-based platforms, IoT solutions in IoT solutions were addressed in some of the
1367 are addressing the significant challenges faced1419
in reviewed papers, including improvements in message
1368 agriculture, such as environmental impact 1420 on routing algorithms and duty cycles in sensor nodes to
1369 communication between sensors, data collection,1421
and reduce power consumption. Additionally, a mission-
1370 processing for improved agricultural management.1422 based approach was utilized to optimize power
1423 consumption in UAVs.
1371 6. Considerations:
1424 7. Future directions
1372 IoT devices use application protocols to transmit data
1373 to cloud-based platforms or applications. These 1425 The future of IoT in smart agriculture is bright. As
1374 protocols, such as HTTP, MQTT, and CoAP, enable 1426 the technology continues to develop, we can expect to
1375 1427
compatibility between non-standardized IoT devices see even more innovative applications that will help
1376 and platforms. For example, Smart Farm 1428 Net farmers to improve their yields, reduce their costs,
1377 implements MQTT and CoAP to allow users to bring 1429 and protect the environment. Some of the potential
1378 their own IoT devices. While HTTP is 1430 not future directions for IoT in smart agriculture include:
1379 specifically for machine-to-machine 1431
(M2M) a) The use of more sophisticated sensors and data
1380 communication, it can be used with REST APIs 1432
to analytics to provide farmers with even more insights
1381 achieve low coupling. However, MQTT is preferred 1433 into their crops and operations.
1382 1434
for smart farming applications due to its resiliency, b) The development of new and improved
1383 interoperability, and transmission rate. Power 1435 communication protocols that are specifically
1384 consumption in IoT devices is highest during 1436 data designed for the challenges of agricultural
1385 1437
transmission, and this review found various methods environments.
1386 for optimizing power consumption in smart farming1438 c) The integration of IoT with other emerging
1387 IoT solutions. These methods include using 1439 low technologies, such as artificial intelligence and
1388 1440
energy protocols, optimizing duty cycles, and using robotics, to create even more powerful and automated
1389 energy-efficient message routing approaches. 1441 farming systems.
1442 V. CONCLUSION:
1390 IoT solutions for smart farming utilize platforms and
1391 cloud computing to store and process data obtained1443 This review paper has delved into the fascinating
1392 by sensors. These large amounts of data can1444 be world of sensors in the Internet of Things (IoT),
1393 analysed using artificial intelligence techniques 1445
like exploring their diverse capabilities and the crucial
1394 machine learning to improve farming management. 1446 role they play in data acquisition. We have examined
1395 1447
For instance, big data analysis can provide insights on the vast array of sensing technologies, ranging from
1396 crops, optimize water usage, prevent disease, and 1448 physical sensors like temperature and pressure gauges
1397 1449
reduce chemical usage. Data collection in both indoor to sophisticated cameras and bio-sensors. Each type
1398 and outdoor environments is carried out using sensors
1450 1493
offers distinct advantages and limitations, making 4. Internet of things for smart factories in industry 4.0, a
1451 them suitable for specific applications. 1494 review. Mohsen Soori, Behrooz Arezoo, Roza Dastres. 26
1495 April 2023
1452 Beyond the hardware, we have also explored the
1496 5. Integration of IoT-Enabled Technologies and Artificial
1453 critical task of data management. The sheer volume 1497 Intelligence (AI) for Smart City Scenario: Recent
1454 and complexity of sensor data necessitate efficient1498 Advancements and Future Trends:
1455 collection, storage, processing, and analysis. We have
1499 Md Eshrat E. Alahil, Arsanchai Sukkuea, Fahmida Wazed
1456 discussed various data management strategies, from 1500 Tina, Anindya Nag, Wattana pong Kurdthongmee, Korakot
1457 edge computing and fog computing to cloud-based 1501 Suwannarat and Subhas Chandra Mukhopadhyay.
1458 solutions, each with its own trade-offs in terms of
1459 scalability, security, and latency. 1502 6. Artificial Intelligence-Based Sensors for Next Generation
1503 IoT Applications: A Review. Subhas Chandra
1460 1504
As we look towards the future of the IoT, several key Mukhopadhyay Sumarga Kumar Sah Tyagi, Nagender
1505 Kumar Suryadevara, Vincenzo Piuri, Fabio Scotti and
1461 trends emerge. Miniaturization and cost reduction 1506 Sherali Zeadally, IEEE. 15 November 2019.
1462 will likely lead to even wider sensor adoption,
1463 enabling denser and more comprehensive data 1507 7. Smart Sensors: Analysis of Different Types of IoT
1464 collection. Artificial intelligence will play 1508 an Sensors. Deepti Sehrawat, Nasib Singh Gill. April 2019.
1509 8. An IoT sensor and scenario survey for data researchers.
1465 increasingly important role in data analysis, 1510 Cleber M. de Morais, Djamel Sadok1 and Judith Kelner. 27
1466 extracting valuable insights and driving intelligent
1467 decision-making. Security and privacy concerns will1511 February 2019.
1468 remain paramount, requiring robust frameworks1512 to 9. An Overview of IoT Sensor Data Processing, Fusion, and
1469 protect sensitive data and ensure responsible use. 1513 Analysis Techniques. Rajalakshmi Krishnamurthy, Adarsh
1514 Kumar, Dhanalakshmi Gopinathan, Anand Nayyar and
1470 In conclusion, sensors are the lifeblood of the IoT, 1515 Basit Qureshi. 26 October 2020
1471 transforming the physical world into a digital tapestry
1472 of interconnected devices. By understanding 1516 the 10. Enabling Artificial Intelligent Virtual Sensors in an IoT
1517 Environment. Georgios Stavropoulos John Violos,
1473 diverse range of sensors and their data management 1518 Stylianos Tsanakas and Aris Leivadeas.
1474 challenges, we can unlock the immense potential of
1475 1519
the IoT to revolutionize various aspects of our lives, 11. Advances in Sensor Technologies in the Era of Smart
1520 Factory and Industry 4.0. Tahera Kalsoom, Naeem Ramzan,
1476 from healthcare and agriculture to smart cities and
1521 Shehzad Ahmed and Masood Ur-Rehman. 27 November
1477 industrial automation. The journey ahead presents 1522 2020.
1478 both exciting possibilities and crucial challenges, but
1479 one thing is certain: the world of sensors in the IoT1523
is 12. How Industry 4.0 and Sensors Can Leverage Product
1524 Design: Opportunities and Challenges. Albérico Travassos
1480 constantly evolving, and its impact on our future1525 is Rosário and Joana Carmo Dias.
1481 only just beginning to unfold.
1526 13. A Systematic Review of IoT Solutions for Smart
1482 REFERENCES: 1527 Farming. Emerson Navarro, Nuno Costa and António
1528 Pereira. 29 July 2020
1483 1. Internet of Things is a revolutionary approach for future
1484 technology enhancement: a review, Sachin Kumar, Prayag, 1529 14. Internet of Things for Smart Cities. Andrea Zanella,
1485 Tiwari and Mikhail Zymbler. 09 December 2019. 1530 Nicola Bui, Angelo Castellani, Lorenzo Vangelista, and
1531 Michele Zorzi. FEBRUARY 2014.
1486 2. A Review of Sensors and Their Application in Internet
1487 of Things (IOT). Anukriti Sharma, Sharad Sharma. March
1532 15. A review of emerging technology for IoT-Based Smart
1488 20213. Review Unleashing the Power of IoT:1533 A City .Md Whaiduzzamn, Alistair Barros, Moumita Chanda
1489 Comprehensive Review of IoT Applications and Future 1534 Supti Barman. 28 Nov, 2022
1490 Prospects in Healthcare, Agriculture, Smart Homes, Smart
1491 Cities, and Industry 4.0. Robin Chataut, Alex 1535
1492 Phoummalayvan and Robert Akl.
1536
1537