Internet of Things
The very name "Internet of Things" indicates the purpose of this emerging technology: to establish a network of Internet connections between all "things", encompassing in the term any element and/or system that surrounds us. In this way, through this technology, from a book to a fridge would be connected to the Internet, which would mean being able to access very useful information about them with a single click.
The concept of IoT was born in 1982, when a Coca-Cola vending machine was connected to the Internet so that it could register the bottles inside and whether they were cold enough. But it was actually in 1999 when Kevin Ashton, while conducting research in the field of networked radio frequency identification (RFID) and sensor technologies at the Massachusetts Institute of Technology, proposed the term "Internet of Things" to describe the system of interconnected devices.
The IoT thus breaks the classical definition of the Internet as the possibility of connecting multiple computers, and takes it further. By breaking the barrier of computers, we add multiple computing devices, mechanical, digital, objects, people and animals to our environment and our network; the only requirement is that such devices must have a unique identifier (UID) and have the ability to transmit data without human or computer intervention.
For example, imagine a smart thermostat. It receives data about the location of your car as you drive so it knows where your car is, and thus adjusts the temperature in your home before you arrive.
You don't have to do anything at all, and it's more efficient and easier than if you had to do it all manually.
Source: https://cdn.pixabay.com/photo/2019/07/02/11/51/zutrittskontrolle-4312260_1280.jpg
The functioning of the IoT has the following components:
- The network: this is nothing more than the network we use frequently, and without which we would not be able to connect to the Internet: Wifi, Mobile Data (3G, 4G and 5G), and Bluetooth.
- The central control system: Once all the data has been collected from the other devices, it is through the network that it is processed until it reaches the central system. It analyses the data to find possible problems and to execute the necessary actions on the different aspects to be considered.
- Devices: these are the different devices or objects that must have a chip with a unique identification, plus the necessary sensors to process the information. These devices can be TVs, watches, vehicles, lamps, medical implants, among others.
But how can IoT be applied to business? IoT can reconfigure facilities and services so that entrepreneurs can keep up with and respond to the needs of a changing market that increasingly demands customised products and services. Similarly, it enables progress on sustainability issues, which is much needed in times of climate emergency.
Source: https://www.youtube.com/watch?v=LlhmzVL5bm8&t=5s (00:00 to 02:52)
A few examples from various sectors can help to understand the changes brought about by the Internet of Things:
Harley Davidson
The production of these motorbikes was very problematic, as it was not efficient and the company's data management was very poor. On top of all this, labour was very expensive, and it usually took a year and a half to implement all the improvements required for the vehicles.
To deal with this situation, they decided to apply the Internet of Things in the New York factory, and in this way, they integrated everything into a single network and all the data could be consolidated.
As a result, it managed to increase its productivity by 80%, reducing the manufacturing cycle from 18 months to 2 weeks and, in addition, improving its profitability by 4%.
Disneyland Orlando: Magicbands
This theme park has become even more magical thanks to the Internet of Things, as it has incorporated wristbands to enhance the park experience.
You can enter the attractions without queuing, you can pay throughout the park, you can locate your relatives, receive the photos you take with your mobile phone, as well as get all kinds of surprises during your stay at Disneyland.
In this way, Disneyland Orlando has managed to offer a unique service to its customers and improve the efficiency of the company.
Virgin Atlantic
Since 2014, the airline's Boeing 787s have been using the Internet of Things in the sky. All parts of the aircraft are connected to the network.
In this way, because all parts are monitored and provide valuable information in real time, flight safety can be increased.
If there is a problem with a part, it will be known before the plane lands. That way, the company can prepare all the necessary repairs before it gets to its destination.
Fitness First
One of the world's largest fitness chains has gone all in on the Internet of Things.
It focuses on customer interaction, as it centres its business model on digitisation to create community among its users and motivate them in their workouts.
Through mobile phones, the company can find out which customers use its facilities and send them information about the gym or about their workouts.
More than just going to the gym, being a customer of this company becomes a unique experience for users and increases the company's efficiency.
Stanley Black and Decker
Stanley Black and Decker is an American manufacturer of industrial and household tools.
The company realised that the productivity of one of its plants in Mexico was decreasing a lot. Therefore, the company decided to use the Internet of Things.
It incorporated radio frequency identification (RFID) devices into the production lines. These devices were responsible for sending information to the system and to the plant managers.
If they encountered problems, employees could call for help by simply clicking a button that sent a message to the connected devices.
Even if they did not report the problem, the devices' own intelligence enabled them to send messages to managers if production efficiency decreased.
Thanks to this implementation of the Internet of Things, the company greatly motivated employees, as they received the help they needed at any given moment.
As you can guess, Artificial Intelligence has the potential to increase the value created by IoT deployments in companies in any sector, to offer new services and operate more efficiently. The IoT is becoming increasingly intelligent, and this fact is not going unnoticed by organisations.
In effect, IoT and AI are two independent technologies that have a significant impact on multiple sectors. Drawing a similarity with our organism, while IoT is the digital nervous system, AI would be a brain that makes the decisions that control the overall system. To refer to the combination of these two technologies, the term Artificial Intelligence of Things (AIoT) has started to be used.
Investments in IoT start-ups using AI have increased significantly and the leading software vendors of IoT platforms now offer integrated AI capabilities. For example, machine learning-based analytics can automatically identify patterns and detect anomalies in the data generated by sensors and smart devices.
IoT projects, linked to Artificial Intelligence, can improve operational efficiency and risk management. In manufacturing processes, predictions from these technologies can reduce the time required by up to 50%, increase uptime and equipment availability by 10-20% and reduce overall maintenance costs by 5-10%. For example, AI-based prediction, especially machine learning, is helping Google reduce 40% of data centre cooling costs.
Many other applications that combine IoT with AI help organisations understand and predict a variety of risks and automate a rapid response, enabling them to better manage worker safety, potential financial losses and cyber-attacks.
For example, Fujitsu has been able, through machine learning, to analyse data from wearables to estimate the risk of heat stress for factory workers. Some banks in India and North America have begun to identify suspicious activity in real time through surveillance cameras attached to ATMs. Even the city of Las Vegas has implemented this new technology to automatically detect and respond to threats in real time.
Robotics
By definition, a robot is a programmable automatic machine capable of performing certain operations autonomously and substituting human beings in some tasks.
As a curiosity, the term "robot" was popularised by the play "Universal Robots Rossum" by Karel Capek in 1920. In the English translation of the play, the Czech word "robota", meaning forced labour or labourer, was translated into English as robot.
The robots we know today were developed after World War II, due to the growing demand for automation in the automotive industry. In the past, robots were no more than tools for automation, programmed to perform specific tasks: transporting, loading, unloading, welding, etc. Today, there are so-called intelligent robots, programmed to detect any disturbance in their environment and to act accordingly.
Source: https://cdn.pixabay.com/photo/2017/08/29/09/46/industry-2692640__480.jpg
A high percentage of jobs are at risk of being automated by robots and this is the major drawback of the implementation of robotic devices in the workplace, as it will cause massive job destruction, mainly in the industrial and service sectors.
It is a reality that is hard to hide. Millions of jobs will disappear around the world in the coming years. But, to compensate, the advent of technologies such as automation will allow millions of new jobs to emerge that were previously unknown. Companies will need to invest in training and retraining for their employees as there will always have to be a human person to control, supervise, check and programme these robots.
Source: https://www.youtube.com/watch?v=htjRUL3neMg (00:27-02:55)
Robots with Artificial Intelligence have long since ceased to be part of science fiction films and have become a reality. Moreover, the combination of robotics and Artificial Intelligence has managed to take a further step that can help companies to develop the digital transformation of their businesses.
Robotic Process Automation or RPA projects that include in their flow some Artificial Intelligence variable are the most complex and evolved automations.
Traditionally, RPA works with structured data and information, i.e. defined in a clear and homogeneous way so that the process is executed without errors. But what if, for example, the information used by the robot does not come from a single source and does not always have the same characteristics? In this case, with classic RPA processes, it would be necessary to study all possible cases and adapt the process to interpret the information in each variable, so that if any new one appeared, it would fail.
With the implementation of Artificial Intelligence techniques, the process itself learns and evolves as new variables come into play. Returning to the previous example, if new data sources were included in a different format than the established one, Artificial Intelligence could make decisions on how to process that information based on the previously analysed information. Thus, thanks to Artificial Intelligence, it is possible to include unstructured data sources in RPA processes.
In relation to the above, we can say that the relationship between robotics and artificial intelligence occurs when Artificial Intelligence is applied to robots, so that they can perform tasks autonomously and, in addition, make decisions on how to do a certain job by applying specific requirements.
Some tasks that, until a few years ago, were performed by people, have been taken over by robots equipped with Artificial Intelligence, so that workers can focus on activities of greater value and creativity for the companies in which they work.
Some of the applications that demonstrate the symbiosis between robotics and artificial intelligence are the following:
Assembly lines. Industry is one of the sectors where robotics combined with Artificial Intelligence has been applied for the longest time. For example, it is quite common to see robots performing automated tasks for assembly or assembly in vehicle factories.
Automated logistics centre. In an automated warehouse, the robots operate autonomously so that, for example, if it is a warehouse of a large supermarket, they can depalletise the shipments of each manufacturer and organise the orders for each shop, always optimising space. This prevents overexertion of workers and accidents at work.
Machinery supervision. Proper maintenance of industrial machinery is one of the keys to optimising results. Robots equipped with Artificial Intelligence can perform this task to improve equipment efficiency and productivity. In this sense, they can perform tasks such as:
? Detect errors in the operation and stop the activity to avoid problems.
? Identify ways to improve the tasks being performed.
? Gathering information to support decision-making.
Packaging. The combination of robotics and Artificial Intelligence is also used in the packaging industry. Robots perform repetitive tasks to pack products, but, in addition, thanks to their sensors and Artificial Intelligence, they are able to make decisions based on the information they obtain.
Robot-assisted surgery. Robot-assisted surgery is a way of performing surgery using tools attached to a robotic arm, so that the surgeon operates the arm with a computer. This allows for great precision and speed when performing complex operations.
Customer service. Many companies in various sectors are already using a combination of robotics and Artificial Intelligence to serve their customers more efficiently. These are chatbots, which are computer programmes that act as tele-operators. The programme provides solutions to users' problems and learns over time. In some cases, the customer service provided by robots is physical, as they are robots that move around and attend directly to the public, for example, in a cafeteria or at a reception desk.
Process automation. The use of robotics and Artificial Intelligence facilitates the automation of tasks and decision-making without the need for human intervention. This does not mean that people are excluded, but rather that robots complement and help.
Word processing and documentation. For example, automation can be used in conjunction with Artificial Intelligence by lawyers to manage the firm's documentation, organise matters and keep clients informed.
Driverless transport. The use of drones and driverless vehicles can be used in a multitude of situations. For example, in the case of a natural disaster to transport medicines to a place that is inaccessible or in the transport of goods. In the case of military transport, it can also be applied, e.g. in the case of drones.
As you have seen, the symbiosis between robotics and Artificial Intelligence opens up a world of possibilities for companies, regardless of their size, in an increasingly changing and competitive environment.
Additive manufacturing
Additive manufacturing consists of converting a digital model into a real three-dimensional solid object without the need for a mould. Instead of using a mould, the material (plastic, metal, resin...) is deposited layer by layer in a controlled manner. The process, also known as 3D printing, starts with a CAD sketch, from which the printing equipment reads the data from the digital model and adds layers of material to form the object.
Source: https://sites.google.com/site/laimpre3d/
Compared to traditional industrial manufacturing techniques, this technology reduces intermediate processes, so parts are manufactured at a higher speed, up to 90% faster. In addition to this, this technique is more sustainable, environmentally friendly and has a lower cost, because only the necessary material is used, no waste is produced. It also saves energy because the parts are lighter, so the machines need less energy to work.
This technology has its beginnings in the 1980s, when Dr Kamoda of the Nagoya Municipal Industrial Research Institute developed a technique for layer-by-layer fabrication of an object.
But it was not until 1986, when Chuck Hull introduced the technique called "Stereolithography", which used liquid resin that solidified under ultraviolet light. Chuck Hull became the founder of 3D Systems with the creation of stereolithography, which marked the beginning of the fourth industrial revolution.
After the invention of this technique, other more advanced techniques were created such as Selective Laser Sintering (SLS), FDM (Fused Deposition Modeling), Binder Jetting, direct laser sintering of metal, material jetting, electron beam melting or DLP (Digital Light Processing) printing.Additive manufacturing is a booming sector due to the aforementioned advantages of speed, precision and cost savings.
It is a technology that has grown rapidly due to its application to rapid prototyping, thanks to which objects with morphological or functional characteristics similar to those of certain products can be created. In this way, designs can be tested as easily as possible before being launched on the market, avoiding errors and guaranteeing the quality of the final products.
However, it is progressively being incorporated as another manufacturing process in sectors such as the textile industry, the automotive industry, or architecture, and there are very relevant examples of its use in medicine, for the creation of personalised implants for each patient, and in the aerospace industry for the manufacture of aircraft engines that withstand temperatures of 700ºC continuously.
Source: https://www.youtube.com/watch?v=EHvO-MlzAIM
However, 3D printing has not yet been able to realise its full potential in terms of productivity. Could Artificial Intelligence be the solution to this problem?
The combination of additive manufacturing and Artificial Intelligence is an opportunity that many experts believe has a bright future.
Currently, Artificial Intelligence facilitates the automation of certain manual tasks, such as data collection, cost tracking and construction planning. Software can also be used to optimise production capacity by improving machine utilisation and scheduling production orders based on availability, all done automatically thanks to Artificial Intelligence. In addition, material selection can also be automated with AI (according to the requirements of the part to be printed, the software makes recommendations on the material to be used to obtain the best result).
In the future, AI will enable the printing of more complex designs, allow the use of new printing materials, and the increased sensing of additive machines will facilitate smarter and more precise control during the printing process.
We can already find examples of progress in this field. At the US company General Electric, they have embarked on the realisation of an intelligent additive manufacturing system. The beginning of their project consisted of a platform to collect data, i.e. all possible information on the manufacture of each part, such as conventional photographs, thermal photographs, scans, analysis of composites and the information generated by each machine. Once collected, they applied machine learning processes to interrelate this data so that the AI can "understand" the entire process. In this way, the AI will learn to determine the errors that may occur and will achieve certain predictive capabilities. This will progressively lead to an intelligent additive manufacturing system that will be able to govern the entire process.
Another example of additive manufacturing coupled with Artificial Intelligence is Astro, the robot dog. Astro was designed and manufactured with 3D printing and has an AI-based brain. It is equipped with optical, sound, gas and radar sensors. It was created to work on security tasks, such as detecting weapons, explosives or weapon residue, but it needs experience to do so. It will learn as it works at it.
Nano and Biotechnology
Nanotechnology is the science of manipulating matter at the atomic and molecular scale to solve problems. At the nano-scale, matter exhibits different properties and phenomena than at the macroscopic scale, because they are governed by the laws of quantum mechanics. For example, they may exhibit new mechanical, optical, chemical, magnetic or electronic properties. It is these new properties that scientists exploit to create new nanotechnological materials and devices.
They achieve this because the properties of materials depend not only on their composition, but also on their structure. Take carbon, for example. The same element, depending on how its atoms are arranged, forms materials as different as carbon, diamond, graphite or graphene.
Source: https://cdn.pixabay.com/photo/2021/05/07/17/03/graphene-6236691__340.jpg
Thus, by manipulating the molecular structure of different materials, we can change their intrinsic properties and obtain other materials with which to implement revolutionary applications in many fields, including engineering, computing and medicine.
As a result, many applications are already available: new electronic components such as MRAM memories, transparent anti-condensation coatings, nanoparticle-based bacterial detection systems, nanodiagnostic and controlled drug release systems, bone substitutes and implants, electrically conductive plastics, dirt- and water-repellent fabrics, others with antimicrobial properties, sun creams, additives for optimising cement-concrete performance, polymer clay nanocomposites for PET recycling, fuel cells, flexible or dual-performance solar cells, anti-friction materials, scratch-resistant paints, anti-wrinkle fabrics, faster microprocessors that consume less energy, batteries that last 10 times longer, etc, etc.
Source: https://www.youtube.com/watch?v=OFV5hqIXSRI
Once again, the combination with AI enables this technology to offer optimal results.
Working at the nanometre scale involves, among many other processes, imaging through electron microscopes capable of viewing nanometric objects. And the data and images obtained by these microscopes need to be processed to enable their observation and interpretation.
This data and image processing can only be carried out by digital means. For example, there are Artificial Intelligence algorithms capable of processing these images, analysing them and detecting atoms in them, while others analyse each of the atoms detected, create digital models of each of them and classify them according to the element to which they belong.
Artificial Intelligence trained specifically for this purpose is able to solve problems such as staggered images, images that have measurements at different heights, or images with a lot of noise or low resolution. In this way, the complex signals obtained by microscopes can finally be transformed into an intuitive image format for users, something that would be impossible for a human to do.
Another common area of work of both technologies, which stands out in particular, is biological processes on a molecular scale.
The activity of the various biomolecules involved in life processes, and the interactions between them, are enormously complex. How can we understand these processes without interfering with them?
Nanotechnology has succeeded in developing nanosensors capable of detecting various compounds without interfering with their activity. These nanosensors can detect the specific "signature" of each molecule, its characteristic way of vibrating in the infrared spectrum.
Artificial Intelligence trained for this purpose is able to distinguish, among the complex stream of data emitted by the nanosensors, the different molecular profiles, based on the known vibration patterns of each molecule. In this way, it can classify the molecules being detected.
This tool can have applications in the field of health, both in pharmacological and medical research and in the diagnosis and development of new therapies and treatments for all kinds of diseases.
This example takes us to other emerging technology: Biotechnology
Biotechnology can be defined as a multidisciplinary scientific discipline, employing biology, chemistry and various processes, with extensive use in agriculture, pharmaceuticals, food science, forestry and medicine, for the purpose of developing new technologies and products.
In fact, biotechnology, understood as any biological modification of other species by humans, has been with us since the appearance of the first agricultural cultures. Today, it is advancing by leaps and bounds and has an increasing number of applications in our daily lives: from pharmaceutical development to food production or the treatment of polluting waste.
Source: https://cdn.pixabay.com/photo/2016/12/13/05/28/dna-1903319__480.jpg
This variety of applications means that there is a classification of the different applications of biotechnology, which are associated with different colours. We highlight some of them, and especially "Golden Biotechnology", the one most directly related to Artificial Intelligence.
Types of biotechnology
Red biotechnology: This corresponds to medical biotechnology, which encompasses all applications related to human health, including the production of new drugs, more effective molecular diagnostics, regenerative therapies or the development of genetic engineering with the aim of finding therapies based on the manipulation of the human genome.
White biotechnology: This is the biotechnology that seeks to optimise the performance of industrial processes through the use of bacteria, yeasts or plants. This increase in yields is achieved thanks to the lower material and energy resource requirements of biological processes.
Green biotechnology: Consists of the treatment and genetic manipulation of plants to improve the processes involved in agriculture. It seeks to increase the production of food, pharmaceuticals, or other products derived from the sector, as well as to reduce negative environmental impacts.
Blue biotechnology: Uses marine organisms both in industry and in the development of health therapies. The development of new products in the food, pharmaceutical and chemical industries are its most interesting applications.
Violet Biotechnology: It deals with the study of the legal aspects surrounding biotechnology as a science: safety measures, patient data protection, patents, bioethical issues and legislation.
Grey biotechnology: Its objective is the conservation and recovery of polluted natural ecosystems through bioremediation processes.
Golden biotechnology: Also known as bioinformatics, it is responsible for obtaining, storing, analysing and separating biological information, especially that relating to DNA and amino acid sequences, and the study of proteins.The applications of Artificial Intelligence in biotechnology are very varied. As mentioned, golden biotechnology is the most directly related to Artificial Intelligence, and a good example of this is Alphafold 2.
Source: https://upload.wikimedia.org/wikipedia/commons/a/a4/C12orf29_AlphaFold.png
Alphafold 2 is an artificial intelligence software trained by deep learning, which is able to make predictions about the structure of proteins.
Proteins are molecules that encompass an overwhelming diversity of functions. Enzymes, hormones, antibodies, storage proteins such as in bird eggs and seeds, transport proteins such as haemoglobin, contractile proteins such as those in muscles, and many other types of structural proteins.
However, their structure always follows the same pattern. They are all made up of a linear sequence of amino acids, which we call the "primary structure". From this, the proteins fold, like a Meccano, in a different way depending on the sequence of amino acids that make them up.
The folding process is very complex, and results in a final three-dimensional structure, which determines the protein we know. This final conformation is of crucial importance, as it determines the biological properties and the function that proteins can perform.
Knowing the DNA sequence of a gene means that we know the primary structure of the protein it will give rise to, because the nucleic acid sequence of the DNA translates into the amino acid sequence. However, it does not allow us to know the protein it will give rise to, because we do not know how the amino acid sequence will be folded.
This is a classic problem in biology, the "protein folding problem". So much so that there is a worldwide competition called CASP (Critical Assessment of Techniques for Protein Structure Prediction) in which different research groups compete against each other to predict the final structure of a "target protein" from the sequence of its amino acids.
The first version of AlphaFold participated in the 13th edition of CASP in 2018, and already took first place in the ranking. The same happened in the 14th edition in 2020 with the AlphaFold2 version, which established the effectiveness of this programme beyond the capabilities of any human in protein prediction.
To put it very simply, we can explain how AlphaFold 2 works: Deepmind, which is the company that developed this application, has turned the problem of protein folding into an image processing problem, an area in which A.I. has a lot of experience. Do you know of any mobile app capable of adding years of age to a photograph of your face? Or capable of changing the face of the Gioconda into your own? AlphaFold uses similar processes to those that make these apps work to convert your photo.
- AlphaFold 2 finds sequences in protein databases that are similar to the "problem sequence", corresponding to genetically close species.
- It makes a comparison between them, and thus locates those amino acids that have changed at the same time during evolution, which means that they are related to each other, close to each other in the structure of the protein.
Source: https://upload.wikimedia.org/wikipedia/commons/6/66/Wikipageimage2.png
- It creates a matrix, an image that expresses the degree of proximity between the amino acids in the "problem sequence".
- The image is processed through a trained algorithm to create a new image, another matrix that constitutes the "pattern" of the target protein.
- The new image is reprocessed by another algorithm that converts the pattern into a three-dimensional design, the structure of the target protein.
Cases such as Alphafold illustrate how Artificial Intelligence has enormous flexibility and capacity to tackle and solve a wide range of problems. In a roundabout way, by translating a molecular biology problem into an image processing problem, a technology company that was not at all related to biotechnology has achieved a huge breakthrough in this field.
This example can serve as an inspiration for solving many other complex problems faced by both the private and public sectors. Artificial Intelligence will undoubtedly bring us many new examples of this "data-driven creativity" in the future.
Source: https://youtu.be/KpedmJdrTpY