The exponential growth in the volume of data requires new ways of analyzing it. In this context, Artificial Intelligence becomes particularly relevant. According to Forbes, the two main trends dominating the tech sector are the Internet of Things (IoT) and artificial intelligence(AI) which creates what we are going to talk about in this article which is the AIOT.
AI and IOT are two distinct and independent technologies that have an important impact on each other. While IoT can be thought of as the digital nervous system, AI would also be an advanced brain that makes decisions that control the overall system.
The concept of the Internet of Things (IoT) has been around for some time, especially in light of our rapid technological development.
The Internet of Things embodies the spirit of physical and digital convergence; the data is collected by an increasing number of devices and then aggregated into what are commonly called “Big Data”. The number of these devices continues to grow, and according to estimates will reach 50 billion by 2020.
The data collected by these devices however runs into a problem when trying to transmit it to a centralized location such as the cloud – specifically, latency.
Despite the connection speed are constantly increasing, they are unable to keep up with the exponential growth in the amount of data. If this problem is not handled, the latency will increase and the overall system performance will suffer.
This is one of the areas in which Artificial Intelligence can make significant contributions. It also paves the way for technological innovations, for example by optimizing city traffic and public safety but also improved financial services.
Internet of Things (IoT) and Artificial Intelligence (AI) are two important information technologies that have now become, over time, undisputed protagonists in various fields of application.
These are two different concepts, but at the same time able to integrate perfectly to create new solutions of value and high potential from a functional point of view.
The use of AIoT requires components capable of managing various conditions and that are at the edge. Its applications can range from onboard vehicles and airplanes to factories or plants in the desert.
All of this requires a flexible and adaptive approach at the same time regarding the production of components. Artificial Intelligence also promises to reduce the human factor at the level of decision making.
This puts pressure on systems integrators to ensure quality control, given that an accident involving Artificial Intelligence, where the human factor is taken away, it will not necessarily have a clear and obvious culprit.
Let’s try to better understand the characteristics of IoT and AI, and how these two realities can be integrated and developed simultaneously.
But before we dig into AIOT, let’s understand and view the whole picture of IOT and AI independently.
Internet of Things: what are we talking about
Internet of Things or IOT, is the ability of various types of devices to communicate data with each other by connecting to the Internet.
Thanks to this technology, the Internet can be extended to the electronic objects we use every day, allowing heterogeneous hardware to communicate and create an ecosystem of devices that transmit data about the world around us through a network infrastructure.
IoT devices combined with artificial intelligence are becoming increasingly popular, thanks also to the widespread use of virtual assistants (Siri, Cortana, Alexa, etc.).
There are now many types of “smart” objects capable of transmitting and receiving data via the Internet: wearable devices, for home automation, for medical monitoring, for transport, for industrial production, agriculture and many others. Before seeing how IoT and artificial intelligence can be combined, we need to understand how IoT technology is structured.
IOT Components
This technology is not independent and in order to use it you need devices with fundamental components:
- Hardware capable of recovering data and capable of executing instructions relating to the object. For example, sensors for temperature or power control.
- Software that manages the collection, storage and manipulation of data collected by the hardware.
- A communication protocol that allows the device to communicate across a network.
IOT architecture
Over the years it has become increasingly efficient and economical to create devices with the fundamental components for using the Internet of Things.
Overall, the architecture of an IoT ecosystem consists of 4 levels:
- Devices (clients)
- Internet gateway (data aggregation system)
- Edge IT
- Cloud
Device layer
The device level includes the options provided by the object. So sensors, hardware actuators and communication protocols used. The sensors will provide input data to analyze real-world information, such as temperature or brightness sensors.
The actuators are used to perform the responses received in a physical action of the device; an example of an actuator is power control, or the ability to turn off a device if certain conditions have occurred.
The communication technologies used are typically of 2 types: wired or wireless.
Typically, Ethernet or Powerline technology is used as wired technologies, which allows the power line to be used also for data communication. As for wireless technologies, some examples of typical use in IoT devices are: Bluetooth, NFC, RFID, WiFi, LTE, etc.
Internet gateway layer
The Internet Gateway is a system that provides a mechanism for processing the data collected by the device sensors and mechanisms for creating a secure connection for data communication. Typically this system aggregates the data provided by the sensors and digitizes them so that they can be transmitted over the internet.
The aggregation mechanism can be implemented both inside the device (as pre-installed software), and in a separate machine near the sensors, which obtains the collected data.
Edge IT layer
This level is considered to be “borderline” (as the name implies), as the data is prepared here and transferred over the Internet. In particular, these systems perform analysis and pre-processing of the data received from the previous level. Machine learning and visualization technologies may be used here, or some additional processing may be done, prior to the data center logon phase.
Given the similarity with the previous phase, it is also possible to combine the 2 levels (Internet gateway and Edge gateway) in a single implementation. In this case, we will obtain that the position of the peripheral IT systems is close to that in which the sensors and actuators are located, creating a unique system of wiring and processing the collected data.
Cloud layer
The Cloud level is the one in which the business rules that coordinate all the devices that are part of the IoT system are implemented. In this level, messages from all devices will be managed and replies to be sent will be created. Storage systems, support applications, and applications for viewing and managing connected devices can be created and managed here.
Applications of IOT
In the future, traffic lights may be equipped with traffic detectors that are activated in the event of queues or delays. This information could be accessible to car navigators, who would recommend alternative routes and divert traffic to less traveled roads.
This would reduce traffic jams and improve the quality of life for both motorists and residents of the busiest areas. Reducing travel times would also save fuel, making cities less polluted.
Cars may also receive information on available parking places over the internet, which saves drivers time and effort. Experiments in this sense are already active in some cities of the United States.
This is just one example of the many functions that the Internet of Things could have within daily life. Within the development of home automation, the IoT applications are manifold.
For example, it is possible that the owner’s car, upon entering the gate, sends an activation signal to the appliances, turning on the lights in the box and starting the cooking of dinner in the microwave, or that the smartphone alarm clock, ringing in the morning, you also activate the coffee machine and the toaster for breakfast.
Interfacing with traffic control, then, the alarm clock could independently decide to ring a little earlier in case of particularly heavy traffic, snow or work in progress on the usual route to work. It is undeniable that the Internet of Things will be able to radically change everyone’s daily life, thanks to increasingly intelligent objects capable of making many decisions on their own.
Internet of Things in the industrial sector There are many possibilities of application also in the industrial and service sectors. The machines could be connected directly to diagnostic devices, reporting any malfunction directly and receiving instructions for self-repair, where possible.
Each object could be monitored at every stage of processing, allowing you to completely trace the path followed and any critical issues encountered, in order to improve efficiency. Already today, in the nascent Industry 4.0, many machines can be controlled remotely.
In animal husbandry, the IoT could be used, as is already partly the case, to monitor individual head of cattle.
Signalers attached to each animal (which in this case functions as a “thing”) would send real-time statistics on the leader’s behavior, productivity, health, and other factors, automatically adjusting food intake, administering antibiotics, and reporting problems that could not be resolved without human intervention.
Plants can also be connected to the network, for example with sensors that regulate their needs for water, light and special fertilization.
The Internet of Things can also revolutionize the automation of an area such as agriculture, radically changing the way crops are managed.
A field managed using smart devices would allow the sector to be completely rationalized, for example by regulating irrigation based on the real needs of the plants.
This would allow significant savings to the 4.0 farm, which would use much less water, and would benefit the environment, since a very precious resource such as water would not be wasted.
IOT issues
Criticality and safety
The many positive aspects that an IoT environment provides have been described, but it is also necessary to mention the possible security risks that this environment entails and what are the main criticalities to set up a system of this type.
Lack of a standard
The first criticality that is encountered in putting together multiple devices to create an IoT environment is the discrepancy in the technical characteristics of the devices.
Since there is still no standard for the creation of such devices, we are dealing with a large number of different hardware and software used, therefore also with a diversity even in communication systems. For example, we may have to let 2 devices communicate using one Bluetooth and the other WiFi.
Privacy and data management
Another issue is privacy and data control. Since data communication takes place via the Internet, we have the same security problems as when we use normal web applications, with the aggravating circumstance that the data passed in the case of the IoT concerns physical devices.
It is therefore important to manage the security / authentication of transmissions since leaving access to physical devices to unauthorized parties can have serious results.
Artificial intelligence (AI)
Artificial intelligence (AI, Artificial Intelligence) is a new evolutionary area of information technology that is based on the idea of creating intelligent and intelligent machines capable of behaving and reacting like humans.
The goal of AI is to help machines operate more humanely by simulating human behavior and intelligence. Intelligent behavior, in turn, involves communication, perception, reasoning, learning, manipulation and action in complex environments.
The goal is to develop machines that can do all of these things as humans do, or maybe even better. Artificial intelligence has both engineering and scientific objectives.
Artificial intelligence systems not only automate industrial processes, making them more efficient, but also allow people and machines to work collaboratively in new ways.
Not just manufacturing, AI systems are being integrated across all major departments, from sales and marketing to customer service to research and development.
The term artificial intelligence expresses the ability of technical devices to simulate the human mind, so that they are able to deal with things as if they were human in a way
Thinking, discovering errors, dealing with matters and handling errors, and carrying out various tasks at high speed and great quality, which allows information to be stored in
Large storage spaces, as when you deal with it, it looks like you are dealing with a professional and experienced person, not an ordinary person with limited intelligence.
We find robots the largest vivid example of the use of artificial intelligence in the machine or technical devices, so that they show a high-performance interaction with us and with great professionalism.
Areas of artificial intelligence
Artificial intelligence is entering into many fields today, as it has become important in many areas that are indispensable, and the most prominent of these areas are:
Medical Field
Which is considered the most field that uses artificial intelligence, due to its importance in it, and its infinite accuracy in performing surgical operations in addition to diagnosing the condition of patients,
Knowing the extent of the patient’s response to treatment, which has come to perform many dangerous and difficult surgeries today, which a person may err in performing.
Industrial Field
It is noteworthy that the role of artificial intelligence is large and wide in the industrial field, especially in the manufacture of machines, because of its accuracy, professionalism and quality in the performance of work, especially
Large and difficult industries, which may endanger humans.
Educational Field
Artificial intelligence also enters the field of education, where the educational process supports and uses artificial intelligence to communicate the idea in a clearer and broader way for students to make good use of it.
Space Field
Artificial intelligence is considered essential in the field of space, from manufacturing space machines and equipment to reaching space, traveling and navigating between planets, etc.
Artificial intelligence features
Artificial intelligence is one of the most prominent things that have been ravaged by the large and modern information and technological revolution in recent years, and all this is due to its unique characteristics.
Many features, most notably the following:
- Artificial intelligence is characterized by thinking in a qualitative and unique way that differs from the way a computer works, as it thinks and deals with the outside world in a way of knowledge, as the human mind deals with the surrounding environment, in addition to the fact that it does not use computer language but rather human languages such as English or Arabic.
- The use of binary symbolic representation, as it does not use digital symbols as a computer does in using numbers, but rather depends on the binary characteristic, which is similar to the way a person understands, thinks, believes and imagines, and therefore he can make decisions based on this binary way of thinking.
- Relying on distinct and unsystematic ways to solve the problems he faces, such as thinking and striving to give solutions. He does not rely on ready-made or pre-designed methods and methods in a systematic way, but rather deals as if you are dealing with a human mind.
The Difference between IOT and AI
The convergence of technologies over the years has blurred the line between the Internet of Things and Artificial Intelligence, but the line is still there.
Looking at what has been listed, we see that the Internet of Things depends entirely on the computer, its algorithms and technologies, and the Internet in all its applications without any interference from.
That said, the IoT is a vast network of interconnected computing devices connected to the Internet. These devices can detect, accumulate and transfer data over a network without any human interaction.
Artificial intelligence, on the other hand, is about creating intelligent and intelligent machines capable of behaving and reacting like humans, providing them with the ability to detect, actuate, store and process data. Artificial intelligence has both engineering and scientific objectives.
To conclude we can say that the Internet of Things (IoT) is a network of connected things that can detect, accumulate and transfer data over a network without any human interaction by providing key physical data and further processing such data in the cloud for provide business information.
Artificial intelligence (AI), on the other hand, is an area of information technology to create machines to do intelligent things as humans do, or perhaps even better.
AIOT: when AI meets the Internet of Things
The Internet of Things is a technology that is helping us in different sectors better in everyday life, but Artificial Intelligence is the real driving force behind the full potential of the Internet of Things, starting from basic applications for tracking our fitness levels, to its wide-ranging capabilities across industries and urban planning, the growing partnership between AI and the Internet of Things means a smarter future.
It will be developing various sectors sooner than we think, integrating artificial intelligence with the Internet of Things will produce the superpowers of innovation:
The Internet is the basis for connecting IoT devices together, to communicate, collect and exchange information about our activities through Internet networks, every day, they generate a billion gigabytes of data, and by 2025, it will be There are approximately 42 billion devices connected to the Internet of Things globally.
Naturally, as the numbers of these devices increase, so will the data sets. This is where AI steps in and brings its learning capabilities to the Internet of Things.
Three main emerging technologies are enabling the Internet of Things (IoT):
- Artificial Intelligence (AI). Programmable functions, systems, and devices that enable devices to learn, infer and process information like humans.
- 5G networks with high speed and near-zero delay for real-time data processing.
- Big Data is where huge amounts of data are processed from many internet-connected sources.
Together, these interconnected devices are changing the way we interact with our devices at home and at work, creating the AIoT
What is exactly AIoT?
Artificial intelligence of things (AIoT) is a general term for applying artificial intelligence to the Internet of Things (IoT), a new phenomenon that represents many simple digital connections between hardware devices.
The artificial intelligence of things represents the technologies that are integrated into the IoT to make it intelligent. AI can add value in helping IoT meet machine learning goals or in using key data to report or develop insights.
The combination of IoT and AI makes it possible to develop increasingly innovative and advanced technological solutions, so much so that we can speak of “AIoT”. But how can these two realities be integrated?
Artificial intelligence exploits a very precious resource in order to function and constantly improve: Big Data! But in order to really talk about an intelligence of machines, similar to that of humans, these resources must be reliable and always available.
To overcome this problem, IoT technologies come into play. The latter are in fact able to collect, aggregate, analyze and provide predictive models by exploiting a very complex system of platforms and devices.
Therefore, the joint use of IoT and AI allows to increase the mutual value of the two solutions: on the one hand, AI increases the potential of the IoT by implementing automatic learning of machines and devices; on the other hand, the IoT increases the value of AI by providing resources in terms of connectivity and data exchange.
The combination of IoT and AI, which we can define as AIoT, therefore allows to obtain even more reliable and precise data, predictive models and functions, providing a solid basis on which to develop new efficient and effective technological solutions.
The use of Artificial Intelligence (AI) technologies with the Internet of Things (IoT) infrastructure to achieve more efficient Internet operations and improve data management and analysis. Artificial intelligence can be used to transform IoT data into useful information to enhance decision-making processes and improve human-machine interactions.
The integration between these two realities can be considered almost essential. The potential of IoT and AI, limited to their unique use, cannot be fully expressed, and only from their union can new advanced platforms take shape that allow the analysis of data and algorithms capable of simplifying the use and operation of devices. smart.
In an ideal future, we could rely on a network of intelligent devices capable of capturing, interrogating, and using data to contextualize and trigger actions in new use cases that enhance our lives.
Why is AIoT needed?
Perhaps the most revolutionary element of AIoT is that it frees us from the traditional cloud architecture we are accustomed to. Rather than putting intelligence in the cloud, and relying on that to interpret data, AIoT makes devices smart in themselves.
Processing and decision-making are done on the machine, allowing for near-instant decisions.
Statista predicts that there will be more than 75 billion IoT devices in 2025, which cloud infrastructures cannot scale to.
As devices begin to proliferate, eliminating latency issues that come with constantly sending data to and from the cloud is essential to the IoT network as a whole.
Beyond that is the question of user privacy: In an age when concerns about consumer privacy continue to grow, data security is more important than ever. Research from Cisco in 2021 found that nearly half (46%) of consumers did not feel they could effectively protect their personal data: Feeling “watched” or “monitored” is a growing concern.
Instead of capturing and transmitting data, AIoT devices can hold the data in devices on board, ensuring that it is not shared with tech giants with potentially nefarious advertising intent. In many cases, these devices do not need to store data at all, which reduces data transfers and makes it more difficult for hackers to access personal information.
This independence is not only digital, but also geographical: AIoT devices have the ability to work almost anywhere.
For example, our current smart devices constantly require stable internet connections for their functionality. AIoT-ready chips enable devices such as smart bulbs or thermostats to operate independently of these networks, providing full functionality even when there are unstable Internet connections.
Given the benefits that AIoT brings to our devices, it is not surprising that many engineers are enthusiastically following its development using the growing demand and opportunities year after year. In fact, the industry is estimated to be worth a staggering $78.3 billion by 2026, according to a Research and Markets report, promising to be one of the most exciting and rapidly growing technology industries out there.
Edge AI, the new trend
Edge AI is a technology that exploits the computing capacity of IoT devices to implement features based on Artificial Intelligence, in order to process and analyze data directly at the source, near the point of acquisition.
Most state-of-the-art artificial intelligence processes run in a cloud as they require a large amount of computing power. As a result, these AI processes can be vulnerable to downtime.
Since Edge AI systems operate on an edge computing device, the necessary data operations can happen locally, being sent when an Internet connection is established, which saves time. Deep learning algorithms can operate on the device itself, the point of origin of the data.
Edge AI is becoming increasingly important due to the fact that more and more devices must use AI in situations where they cannot access the cloud. Consider how many factory robots or how many cars today are equipped with computer vision algorithms.
A delay in data transmission in these situations could be catastrophic. Self-driving cars cannot suffer from latency when detecting objects on the road.
Since a fast response time is so important, the device itself must have an Edge AI system that allows it to analyze and classify images without relying on a cloud connection. When edge computers are entrusted with the information processing tasks normally performed in the cloud, the result is low-latency, low-latency real-time processing.
Furthermore, by limiting data transmission to only the most vital information, the data volume itself can be reduced and communication interruptions can be minimized.
Many companies are, at the same time, looking for AI solutions in the Cloud, but centralized resources, although unsurpassed from the point of view of computational resources and scalability according to workloads, are currently unable to answer all of them. the operational needs of a digital business.
The issue is now widely known. The Cloud provides large computing and memory resources, but the distance from the data can generate objective problems of operational latency, together with the conditions in which an Internet connection is not available to allow the exchange of data between centralized and peripheral systems.
The IoT devices, placed in peripheral conditions and interconnected through the edge architectures, act in proximity to the data acquisition point, in conditions of reduced latency, but are not able to satisfy the most demanding computing operations. This functional context also applies to applications based on Artificial Intelligence.
As for the contribution of various Artificial Intelligence technologies, Edge AI can offer efficient solutions for low-latency real-time processing, while Cloud AI can be used for the most computationally demanding workloads.
such as operations based on Deep Learning, which require enormous amounts of data and computational resources to complete the learning and analysis processes for which they are developed.
An IoT system equipped with Edge AI functionality can in fact operate selectively on the data, skimming at the source the information to be transmitted to the Cloud AI services, which coincide with those actually required by the analyzes.
This approach makes it possible to significantly optimize the use of bandwidth and storage and significantly reduce costs, considering that Cloud services are almost always based on a pay-per-use pricing model that can be scaled according to the resources actually used.
Realistically, although IoT devices are increasingly powerful and performing, especially as regards the more complex operations, such as those in the industrial sector, it is currently difficult to think of a fully autonomous, functionally, Edge AI system.
The availability of some resources typically accessible in a centralized data center must be functionally contemplated for the needs to be met, avoiding radical solutions in the technological approach.
The key to success therefore lies in finding the correct balance, creating the right mix between the speed and connectivity of the Edge.
with the processing power and storage capacity of the Cloud, without neglecting the aspects related to security and compliance with the regulations relating to data processing, which often condition the choice of an IT architecture even before the reasons of technological opportunity.
Edge AI and the Internet of Things
Edge AI integrates with other digital technologies such as 5G and the Internet of Things (IoT). The IoT can generate data for Edge AI systems to use, while 5G technology is essential for the continued advancement of both Edge AI and the IoT.
The Internet of Things refers to a variety of smart devices connected to each other via the Internet. All of these devices generate data, which can be fed into the Edge AI device, which can also act as a temporary storage unit for data until it is synced with the cloud. The data processing method allows for greater flexibility.
The fifth generation of the mobile network, 5G, is critical to the development of both Edge AI and the Internet of Things. 5G is capable of transferring data at much higher speeds, up to 20Gbps, while 4G is capable of delivering data at only 1Gbps.
5G also supports far more simultaneous connections than 4G (1,000,000 per square kilometer versus 100,000) and better latency rates (1ms versus 10ms).
These advantages over 4G are important because with the growth of the IoT, the volume of data also grows and the transfer speed suffers. 5G allows for more interactions between a wider range of devices, many of which can be equipped with Edge AI.
Use cases for Edge AI
Use cases for Edge AI include virtually any instance where data processing would be performed more efficiently on a local device than when done via a cloud. However, some of the most common use cases of onboard AI include self-driving cars, autonomous drones, facial recognition, and digital assistants.
One of the most important use cases for Edge AI is self-driving cars. Self-driving cars must constantly scan the surrounding environment and assess the situation, making corrections to its trajectory based on nearby events.
For such cases, real-time data processing is important and, as a result, their integrated Edge AI systems are responsible for data processing (storage, manipulation and analysis). Edge AI systems are required to bring Tier 3 and Tier 4 (fully autonomous) vehicles to market.
Since autonomous drones are not piloted by human operators, they have very similar requirements for autonomous cars. If a drone loses control or malfunctions during flight, it can crash and damage property or life.
Drones can fly out of range of an internet access point and must have Edge AI capabilities. Edge AI systems will be necessary and crucial for services like Amazon Prime Air, which seeks to deliver packages via drone.
facial recognition systems is another use case of Edge AI. Facial recognition systems are based on computer vision algorithms, analyzing the data collected by the camera. Facial recognition apps that operate for the purposes of tasks such as security must function reliably even if they are not connected to a cloud.
Another common application of Edge AI is digital assistants. Digital assistants like Google Assistant, Alexa, and Siri need to be able to operate on smartphones and other digital devices even when not connected to the internet. When data is processed on the device, it does not need to be sent to the cloud, which helps reduce traffic and ensure privacy.
As the adoption of NN-based algorithms generally provides significant benefits, and although in recent years these algorithms have been primarily defined to run in the cloud, lately a promising new trend is emerging, called EdgeAI and related to the transfer of intelligence from the cloud to at the edge.
In fact, the execution of AI algorithms on IoT edge devices that are close to the data sources (i.e., the IoT nodes) offers undoubted advantages, such as reducing the amount of data to be forwarded from the edge to the cloud, thus reducing network load and latency, and supporting scalability.
The limits
As a disadvantage, however, AI algorithms running on-the-edge must be designed taking into account the limitations imposed by IoT devices, mainly in terms of required memory, computational and energy resources, significantly lower than those offered by cloud platforms.
Therefore, when modeling EdgeAI algorithms, it is necessary to seek a balanced compromise between
- The will to achieve the best performance (in terms of prediction and prediction) and
- The implementation of a prediction AI model “Light enough” at the computational level, so that it can be run by an IoT device.
In this regard, to quantify this trade-off, various performance indices exist in the literature that can be used as an estimate of the decision.
How will AIoT be implemented?
While the potential of AIoT is clear, many of our current devices limit their realization. There is a wonderful art to balancing power, cost, and flexibility in designs, which often results in a flow of innovation.
Research conducted at XMOS shows that processing power issues are a constant barrier in the minds of engineers, with more than half (53%) citing lack of power as a major issue.
This is not because power is not available but because it is expensive and not necessarily designed for a specific purpose – it can be nearly impossible to find the right devices to strike the right balance in the power/cost ratio.
As a result, in the 4th century CE, 48% of engineers cited costs as a major obstacle to AIoT adoption, with 73% saying the prices of the devices were already too high. Existing chips that pack enough punch to take advantage of AIoT are simply too expensive, big, and solid — a poor match for the many smaller devices that will make the most of the technology.
However, while the barriers are obvious, they can be overcome.
We’re starting to see the emergence of processors designed specifically for AIoT that balance AI, DSP, control, and I/O .
this happens in a way that allows the product designer to define them in software, rather than a chip vendor who installs them in hardware, and do so in a cost-effective package suitable for mass production.
AIoT Key Segments:
So where are AI and IoT heading together?
The joint use of Internet of Things tools and artificial intelligence algorithms (AIoT) allows you to create synergy between worlds that represent the future of technological evolution and enable very heterogeneous scenarios. Here are some examples of uses in which the AIoT represents an interesting integration.
Wearable devices: All wearable devices include the performance of important tasks, such as smart watches, they constantly monitor and track user preferences and habits.
Not only have these applications had an important impact on the health technology sector, but they are also working effectively and are encouraging sports and fitness, according to leading technology research company (Gartner), where the global wearable device market is expected to see revenue of more than $87 billion by year. 2023.
Smart homes: homes that fulfill all your requirements and luxuries are no longer limited to science fiction, where homes can gradually turn into smart homes, achieved by making use of devices, lighting and electronic devices connected to your smartphone, and where these devices will learn the habits of the home owner and develop Among its automated functions and features,
this seamless access also provides additional benefits to improve energy efficiency. As a result; The smart home market could witness a compound annual growth rate (CAGR) of 25% between 2020-2025, reaching $246 billion.
Smart Cities: With the increasing trend of many residents of rural areas to choose urban areas to live in, as cities provide continuous development, to achieve safer and more convenient places to live. Smart city innovations are aligned with investments directed towards improving public safety, transportation, energy efficiency and improving productivity.
The practical applications of artificial intelligence as the ultimate means of traffic control are already becoming clear. In New Delhi, India, one of the world’s busiest cities, an Intelligent Transportation Management System (ITMS) is used to make dynamic, real-time decisions about traffic flows.
Smart industries: When talking about industries and their development, it is necessary to focus on the matters on which these advanced industries depend, as industries from manufacturing to mining depend on digital transformation to improve efficiency and decrease human error. From real-time data analytics to supply chain sensors, smart devices help prevent costly industry errors. In fact, Gartner also estimates that more than 80% of enterprise IoT projects will integrate AI by 2022.
Thus, the integration of AI into the Internet of Things has become more extensive; It continues to push the boundaries of data processing, intelligent learning, and machine learning technologies.
Where it will be the same as before as any company that happily ignored the Internet at the turn of the century, companies that rejected the Internet of things risk being left behind.
Despite being a relatively new technology, artificial intelligence seems to have been around for years. Despite the hype, the functional side of AI is certainly not exciting. Mostly found in remote data centers, they are capable of processing huge amounts of information using complex algorithms – with varying degrees of autonomy.
The exciting part is the benefits that AI brings to both business and consumers: time it can save, tasks it can automate or eliminate, safety improvements, and impressions. AI is an invaluable tool for many businesses while simultaneously establishing itself as a part of our non-working lives – be it in our phones, Google searches, or smart homes.
Examples of synergy in details
An example of synergy is also of interest to the European Union, which within the Horizon 2020 (H2020) program finances, among others,
Furthermore, this “technological joint venture” also favors the migration of AI from centralized cloud computing infrastructures towards the more innovative paradigm which is edge computing.
This push of AI towards the edge computing paradigm is also observable through the recent announcements in the world of consumer electronics.
Global ICT giants are increasingly reducing the prices of their speech recognition models to allow use also on embedded devices in “isolated” mode without the need for a stable and persistent connection to the Internet.
and of its recently released hardware products – an example is Google, which in October 2018 released a processor called Edge TPU, specifically defined to run specific AI TensorFlow Lite models on edge-type devices.
As anticipated, the joint use of AI and IoT is an enabling element for very heterogeneous scenarios: below, by example, are some uses in which the IoT represents an interesting integration.
Management of autonomous vehicle fleets
The IoT can also be used for the management of vehicle fleets, for the monitoring of vehicles in terms of planned maintenance and, as far as possible, also with a predictive approach, to reduce the costs of the fuel necessary for individual vehicles, and also to identify abnormal or unsafe behavior of drivers.
The use of IoT devices such as GPS and other sensors, together with an AI system capable of processing the data arriving from the various components of the system, favors the AIoT even in these contexts.
The application of AIoT-oriented technologies for driving safety is a topic of particular interest to both the academic world and industry, and the European Union also favors research in this direction.
An example is related to the European project “Next Generation Smart Perception Sensors and Distributed Intelligence for Proactive Human Monitoring in Health, Wellbeing, and Automotive Systems”.
The goal of the NextPerception project is precisely to combine the development of innovative sensors with distributed intelligence techniques to build people monitoring services, even inside vehicles, while driving.
Another use of the AIoT concept is related to autonomous vehicles, such as automatic driving systems present on board cars, and defined by various companies (e.g. Amazon, Apple, Audi, Baidu, Ford, Hyundai, Microsoft, Nvidia , Samsung, Tesla, Toyota, Uber, Waymo), which use radar, sonar, GPS and cameras to collect data on driving conditions and, coupled with an AI system, make decisions on any actions to be taken in situations of danger.
Monitoring of large areas using flying drones
Another example of the application of the IoT concerns the monitoring of traffic in a smart city using drones.
Real-time control, with the consequent possibility of requesting changes to the flow of traffic (e.g. by updating the on / off policies of traffic lights), represents an innovative possibility also in terms of environmental sustainability, in order to reduce the urban congestion rate.
In detail, when drones are used to monitor a large area, they are able to transmit data on monitored traffic, which are used by AI models to perform various types of analysis and make decisions to safeguard the situation.
Examples of such models are like detection of accidents and parking situations in disallowed zones, i.e. the modification of the activation sequence of a certain series of traffic lights to allow emergency vehicles to reach patients in need of health care more quickly.
Also in this area, the European Union is supporting research by Member States, financing innovative projects such as, for example, the “Airborne Data Collection on Resilient System Architectures” (ADACORSA) project.
The project aims to use IoT technologies to improve the effectiveness and safety of drones, with particular attention to operational scenarios beyond the line of visibility (Beyond Visual Line of Sight, BVLOS), to allow drones to perform complex operations and move around while minimizing human intervention.
Smart farming
The introduction of heterogeneous technologies (such as those related to the IoT and AI) in the agricultural sector, with the aim of improving productivity and sustainability, is currently a consolidated practice.
In fact, thanks to this trend, in recent years the production processes of horticultural greenhouses have been optimized, for example in terms of increasing automation, tending towards the concepts of Smart Agriculture or Smart Farming.
To this end, greenhouses play a crucial role in agricultural production, since, by defining optimal growth conditions for indoor crops, vegetables, fruit, herbs and other types of edible products can be grown at any time and anywhere. regardless of their seasonality or the (possibly adverse) meteorological conditions of their growing area.
In the context of greenhouse crops, the IoT can represent an enabling factor especially for the definition of methodologies for the development and maintenance of an adequate growth habitat (also defined as “microclimate”),
corresponding in detail to a complex set of environmental variables inside the greenhouse (e.g. soil humidity, air temperature and humidity, level of solar radiation in different points of the greenhouse, etc.)
which, if examined in an interconnected way, allow a precise analysis of the possible actions to be carried out inside the greenhouse (eg cooling, heating or ventilation systems installed internally), whose data could also potentially be correlated to external factors (eg weather conditions, wind speed, etc.).
With reference to the simplified and automated monitoring made possible by the adoption of AIoT technologies in Smart Agriculture contexts, it is possible to identify three main areas of intervention.
1- Values (sensory) related to relevant environmental variables inside the greenhouse, which must be kept within well-defined adequate ranges (for example, air humidity and temperature), can be collected through IoT devices equipped with sensors generally organized as wireless networks of nodes sensory (Wireless Sensor Network, WSN).
Furthermore, given the structural characteristics of the IoT nodes, which require low energy consumption and a longevity of the order of years, usually the data collected by these nodes are sent to less constrained entities, often referred to as gateways (GW) and connected to the Internet.
Hence, GWs generally forward IoT node data to processing and computing infrastructures located within cloud computing platforms to be subsequently retrieved and visualized for the benefit of both end users (farmers) and industry researchers.
2– Within such a collection system, additional control devices (eg actuator nodes) can be integrated, installed inside the greenhouse to regulate the internal climate. For example, if an IoT sensory node detects a dangerous air humidity index, it would be possible to automatically activate a ventilation system to lower the humidity of the air.
3– In support of data acquisition and implementation, the development of complex models and predictive algorithms, with the aim of predicting future values of the monitored environmental variables, represents a vanguard in the agricultural sector.
By way of example, the internal variables of a greenhouse monitored by IoT nodes could be predicted in a satisfactory way through Deep Learning (DL) algorithms based on Neural Networks (NN).
This could also benefit situations in which there are data shortages at certain times: using AI algorithms, it would be possible to deduce the missing sensor data, such as those not correctly collected inside the greenhouse due to a temporary lack of network connectivity, as well as for maintenance operations.
Buildings and commercial establishments
A first example of application of the IoT can be linked to buildings with intelligent offices, equipped with a network of environmental sensors able to detect the staff present and consequently adjust the environmental parameters (for example, temperature and lighting) to improve the energy efficiency of individual offices and of the entire commercial building.
Another aspect that can benefit from the IoT may relate to the control of accesses to the intelligent building through facial recognition technology for visitors and employees.
The combination of connected cameras, whose output images are analyzed by means of AI models that can compare the images taken in real time with a database to determine who should be allowed to enter the building, represents a perfect example of the application of the concept of the IoT.
Similarly, employees would not need to stamp entry or exit from the building, and this would also contribute to an improvement in workplace safety policies in terms of unauthorized attendance in particular scenarios.
In the case of commercial businesses, such an AIoT system could help in the recognition of customers when they enter the business.
At this juncture, the system would be able to collect information on customers, including preferences on products of interest, the flow followed within the store, while analyzing the data to accurately predict customer behavior, and using this information to make decisions on commercial strategies to be taken within the shop (e.g. marketing, targeted advertising for age groups, positioning of certain products in certain strategic points of the store, etc.).
Companies investing in AIOT
Bosch aims to leverage the competitive advantage of its extensive experience in combining connectivity (Internet of Things, IoT) and artificial intelligence (AI) to generate new business and become an AIoT leader. Bosch expects AI-enabled products to generate billions of euros in revenue over the next few years.
Sales of connected devices for the home are expected to double from four million units last year to around eight million in 2021.
In addition, Bosch intends to use AI to evaluate data on how customers use. products in order to provide software updates to create new features and services for customers.
Amazon: AWS IoT connects with AI services, allowing us to make smarter devices even when they are not connected to the Internet.
According to Amazon, “AWS IoT delivers broad and deep capability, spanning the edge to the cloud,” allowing developers to create IoT solutions for “almost any use case across a large range of devices.”
Nvidia has launched a new division called Autonomous Machines, which includes the JETSON XAVIER NX, the World’s Smallest AI Supercomputer for Embedded and Edge Systems, which is dedicated to AIoT solutions.
Xiaomi is investing heavily in of smart homes, it plans to spend 1.5 billion dollars a year on the development of artificial intelligence research and 5G technology.
Is it worth it?
Against Covid
Artificial intelligence applied to the IoT can also help companies and institutions manage that social distancing which, in the midst of the pandemic, has in many cases also become a legal obligation.
Intelligent cameras with corporate AI can report attendance data on a single platform and therefore access to shopping centers, stations or workplaces is allowed only when it can be done in real safety. Indeed, before the situation becomes critical, it is possible to set alert thresholds and also to examine the flow histories to alert customers and prevent them from accessing the shop or supermarket.
Not a secondary detail, intelligence allows the data provided by the cameras to be made completely anonymous, ensuring that people’s privacy is always protected.
Deep learning
Going well beyond this pandemic, which sooner or later, we all hope, will end, the intelligent analyzes of deep learning, to stay with video surveillance, as far as object detection is concerned, they offer security officers greater awareness of the situation and better possibility of verification, thus allowing an immediate response to any threats. It is said, and it is true, that metadata lowers detection times from hours to minutes, if not seconds.
Beyond Covid
Even specialized and tailor-made solutions for virus containment, such as thermal imaging cameras or mask detection, if they fully exploit artificial intelligence, can, in a very near future, we hope, be used in other less “emergency” areas.
Sensors, cameras, network infrastructure, big data, cloud, all the components of artificial intelligence technology, once in the field, can easily be directed towards other functions: the whole heating unit, for example, for fire prevention or for access control of a person in a building.
Also because the development of 5G, experts tell us, will allow the AIoT to extend from smart home or commercial applications to those at an industrial level.
Industrial IoT
Industrial IoT, in fact, makes processes efficient, productive and innovative by enabling an architecture that provides real-time information on operating and business systems. The data is converted into instructions that enable machines to perform specific tasks.
The system based on artificial intelligence takes less time and can run continuously without errors. As a result, production efficiency improves and takes much less time. Of course, the key will always be the flexibility of the devices.
An interesting market
In essence, it is a market in which it is worth investing. Recent research by Markets and Markets started from the very basics, the chips.
The size of the global artificial intelligence market in this sector was $ 7.6 billion in 2020 (despite the virus …) and is expected to reach $ 57.8 billion by 2026, with a CAGR of 40.1%. The main drivers for the market are the increasingly large and complex data, the growing adoption of deep learning and neural networks.
These are those networks that, as the name implies, are modeled on the basis of the functioning of our brain, certainly the most intelligent system that has ever been conceived. It is impossible not to focus on these data and on these technologies.
The Future of AIOT
These devices pave the way for AIoT to enter the mainstream, and will revolutionize the technology industry as we know it.
The smart home introduces new functions every year, to work towards complete and seamless control of our surroundings. Cameras with face recognition can easily distinguish between family and friends — and the ability to recognize unwanted visitors or identify a vulnerable relative in trouble. In time, these devices will talk to each other: an unwanted face in the back door can slam your windows.
You can take this logic outside of the smart home into the world of healthcare for example. Imagine devices that can monitor your breathing and heart rate, able to alert emergency services or share key findings or patterns with your doctor, without revealing any of that data to those who might use it for advertising.
Ultimately, once manufacturers address the barriers to widespread adoption, we will begin to see the wave of innovation that will herald the creation of AIoT into the mainstream. It is destined to become a multi-billion-dollar industry along with the internet, computer and smartphone as one of the biggest technological breakthroughs in modern history.