The impact of Artificial Intelligence (AI) is highly recognized as a key driver of the industrial digital revolution together with data and robotics 1 2. To increase AI deployment that is practically and economically feasible in industrial sectors, we need AI applications with more simplified interfaces, without requiring highly skilled workforce but exhibiting longer useful life and requiring less specialized maintenance (e.g. data labelling, training, validation…)
Achieving an effective deployment of trustworthy AI technologies within process indsutries needs a coherent understanding of how these different technologies complement and interact with each other in the context of domain-specific requirements that industrial sectors require3, such as process industries who must leverage the potential of innovation driven by digital transformation, as a key enabler for reaching Green Deal objectives and expected twin green and digital transition needed for a full evolution towards circular economy.
One of the most important challenges for developing innovative solutions in the process industry is the complexity, instability and unpredictability of their processes and impact into their value chains. These solutions usually require: running in harsh conditions, under changes in the values of process parameters, missing a consistent monitoring/measurement of some parameters important for analysing process behaviour and difficult to measure in real time. Sometimes, such parameters are only available through quality control laboratory analysis that are responsible to get the traceability of origin and quality of feedstocks, materials and products.
For AI-based applications, these are even more critical constraints, since AI requires (usually) a considerable amount of high-quality data to ensure the performance of the learning process (in terms of precision and efficiency). Moreover, getting high quality data usually requires an intensive involvement of human experts for curating (or even creating) the data in a time-consuming process. In addition, a supervised learning process requires labelling/classifying the training examples by domain experts, which makes an AI solution not cost-effective.
Minimizing (as much as possible) human involvement in the AI creation loop implies some fundamental changes in the organizations of the AI process/life-cycle, especially from the point of view of achieving a more autonomous AI, which leads to the concept of self-X AI4 . To achieve such autonomous behaviour for any kind of application it usually needs to exhibit advanced (self-X) abilities like the ones proposed for the autonomic computing (AC)5:
Self-X Autonomic Computing abilities
Self-Configuration (for easier integration of new systems for change adaptation)
Self-Optimization (automatic resource control for optimal functioning)
Self-Healing (detection, diagnose and repair for error correction)
Self-Protection (identification and protection from attacks in a proactive manner)
AutonomicComputing paradigm can support many AI tasks with an appropiate management, as already reported in the scientific community 6 7 . In AI acts as the intelligent processing system and the autonomic manager (continuously executes a loop of monitoring-analyzing-planning-executing based on the knowledge (MAPE-K) of the AI system under control for developing a self-improving AI application.
Indeed, such new (self-X) AI applications will be, to some extent, self-managed to improve their own performance incrementally5. This will be realized by an adaptation loop, which enables “learning by doing” using MAPE-K model and self-X abilities as proposed by autonomic computing. The improvement process should be based on continuous self-Optimization ability (e.g. hyper-parameter tuning in Machine Learning). Moreover, in the case of having some problems in the functioning of an AI component, the autonomic manager should activate self-Configuration (e.g. choice of AI method), self-Healing (e.g. detecting model drify) and self-Protection abilities (e.g. generating artificial data to improve trained models) as needed, based on knowledge from AI system.
In just a few weeks, CARTIF will start a project with the help of AI experts and leading companies of various process industry sectors across Europe to tackle these challenges and close the gap between the AI and automation by proposing a novel approach for a continuous update of AI applications with minimal human expert intervention, based on an AI data pipeline, which exposes autonomic computing (self-X) abilities, so called self-X AI. The main idea is to enable the continuous update of AI applications by integrating industrial data from physical world with reduced human intervention.
We’ll let you know in future posts about our progress with this new generation of self-improving AI applications for the industry.
From the smartphone we carry every day, the tablet or the computer, till any other portable electric tool we use in our everyday have implicit the use of an electric energy accumulation system, or what is commonly known as batteries, in this case rechargable batteries.
But, we really know what batteries are, what contain or how the materials that make them function can be recovery?
Many times the unknowledge of our environment make us carrying a bad management of some of the elements that surrounds us when they reach their service life.
Before knowing these details, could you tell me how many types of batteries exists nowadays?, we talk about Nickel Metal Hydride, Nikel Cadmium or we focus on lithium-ion, now on everyones´ lips?
Nikel Cadmium are used mainly to feed computers, mobile phones and wireless and some varieties of toys, but they are used less and less.
Nikel Metal Hydride are a battery variety less harmful for the environment and with a longer service life.
Lithium-ion are the batteries with the biggest energy storage capacity in comparison with the previous ones and those that are currently most widely used.
Although this post could go on for as long as some of the encyclopaedias have volumes, those that gather dust on our shelves at home, the initial idea is to get to know lithium-ion batteries a little better and why is necessary to attend the recovery of its materials at the end of their service life.
To understand the importance of this need for materials, it is necessary to understand the dependence of our European continent on raw materials, critical raw materials such as the ones that we found in nowadays Lithium-ion batteries as cobalt, nikel, lithium or manganese. Much of these materials are concentrate in very specific places of the planet, which creates a greater dependence on these.
Right, we already know that exists different types of materials inside lithium-ion batteries, but let´s make it a little more complicated, so it not only exists one type of lithium-ion battery, but, depending on its application, we talk about different chemicals, that is to say, the components that form the different cells of the batteries are based in different materials, quantities and conglomerate, as well as different morphologies. These different, lets say models, are changing since their invention at the end of the 90´s, because of their dependence on raw materials or because of the technological advances. We can count with up to 6 different types of lithium-ion batteries models. And in case you were thinking about it,yes, this will complicate their recycling.
We have already assume that we are dependent in terms of raw materials, but, in addition, we have to add the tendence to decarbonization of our energetic system, that mainly at the transport sector is tending to electric vehicle, that as we already know, uses lithium-ion batteries. Europe´s goal is to achieve carbon neutrality by 2050.
Going back to the initial question, we already know which materials make up a battery and that there are many types of them, but in addition we know the need of our european community in terms of reuse of these materials, therefore, we would have to recover those materials at the end of the lithium-ion batteries life service, but, how it is done?
Currently it exists 3 huge methods for recycling those batteries named pyrometallurgy, hydrometallurgy and direct recycling, whose influence over the value chain is next one:
Pyrometallurgy: high temperature foundry process, it should be made up of 2 steps: first, batteries are burnt in a foundry, where the compounds are decomposed and organic materials are burnt, such as the plastic and the separator; the new alloys are generated by the ashes carbon reduction.
Hydrometallurgy: in this process, the materials recovery is achieve by an aqueous chemistry, through the leaching in acid disolutions (or basic) and his later concentration and purification, by the evaporation or separation of the solvent. Purity and quality of the extracted metals are usually differentiated according to this last purification stage of the process.
Direct recycling: recovery method proposed for reaconditioning and recover directly batteries active materials, preserving their oirginal structure.
If we pay attention to carbon neutrality, the first method will no longer be feasible at long term, so involves a series of green house efect emissions associated, therefore the most sustainable ways would be hydrometallurgy and direct recycling.
You thought it would never happen, but you´re watching it happen. Your world upsidedown at an unexpected speed. Ecologists announced a different world according to their believes, but it turns out that in the end it will be the cold sceptics of the Excel sheet who will do it. Ukraine war has caused an energetic crisis, and we wil se if it won´t also be food, that it doesn´t only brings us high energy prices, but also could cause shortage of gas, petroleum and offshots.
We are seeing that in order to resolve this situation it is being proposed to tap into Europe´s subsoil resources, especially shale gas, and to increase generation capcity based on nuclear fission. All these measures could serve to alleviate the energy crisis, although it does not seem at this stage to be willing to disengage from greenhouse gas and pollutant emissions. So it is likely that we will not see much hydraulic breakup, we will probably see more nuclear reactors and, above all, we may see a strengthening of the energy efficiency and renewable generation policies that the European Union has been promoting for some time. And it will not be for environmental reasons, but simply to maintain an economic system that does not take us back to the 18th century.
The sun and its child, the wind, will increase their weight in the electric system faster than expected if access to the raw materials needed to manufacture generators is not interrupted. The stoarge of energy could be developed with intensity and we end up getting acquainted with hydrogen as we have made in the past with butane. But surely what we have the hardest time getting usd to would be the new figures that will appear in the energy system management.
The energy communities are one of the news that are getting shape in Spain. Although still aren´t frequent, there are several examples of people that joint to generate and manage the energy they consume. The downgrading of the photovoltaic panels favours their installation in domestic roofs, which achieves that generation and consumption are close. Energy management could be done from the cloud thansk to Internet of Things and specialized companies could offer this service to communities. Hydrogen and batteries seems to be called to be the energy storage medium, although it will depend on the cost and availability of raw materials. Internet of Things woul allow to manage demand flexibility inside the community. It seems to start being possible that a group more or less big of citizens constitute their own electricity generation company.
But for these participative companies, this capitalism at a human scale, could be possible, we have to defeat some obstacles. And leaving aside reluctance to change, the mosr important is the cost of setting up such a community. Are being made huge efforts to understand people motivations1 to get involved in an energy community and to design mechanisms to set them in motion2, but perhaps not as much effort is being put into designing the business models that would make them economically viable.
We can think of some business models for energy communities. The most clear is the save in energy purchase. If the community generates their own energy and distributes it betweent their members, they will save at least the trasnport tolls that are payed in a conventional bill. Other possible business would be the sale of energy surplus, but current legislation imposes limitations on the distance at which the buyer can be located. The demand flexibility could also give rise to another businees model based on promote a distribution grid of auxiliary services, but this is not easy. If this were to be attempted through balancing markets, the regulations impose minimum power values that will be difficult for many communities to achieve. Moreover, it should be borne in mind that it is not possible to interact with the network without complying with a whole series of complex technical rules. It becomes necessary the independent aggregator figure, which is already provided for in existing legislation, but which is not fully developed and which would have to intermediate between the community and the electricity grid. These problems could be solve if they existed energy local markets or flexibility markets, but in Spain are in an embryonic state and it will still take some time to see them in operation.
But, despite of these deficiencies, nowadays energetic crisis overview joint with the directives that came from the European Union will boost the development of energy communities. The problem will be finding resources to do so. Administrations and the cold sceptics of Excel spreadsheets who come up with innovative business models may have the last word.
Cultural and Natural Heritage (CNH) are irreplaceable sources of life and inspiration, according to the UNESCO definition. Europe´s rural areas represent outstanding examples of cultural, either tangible or intangible, and antural heritage that need, not only to be safeguarded, but also promoted as an engine for competitiveness, growth and sustainable and inclusive deveopment1. According to the PAHIS 2020 Plan2 , there has been a deepening of the so-called Cultural Heritage Economy in recent years, in accordance with current criteria which establish that cultural heritage assets should no longer be perceived as a burden but as a resource capable of generating development and social cohesion. This post gives a brief summary into the study of computer technologies applied to modelling and monitoring how the CNH can support the sustainable development of rural areas.
The EU communication “A Long-Term Vision for the EU´s Rural Areas”3 mentions the EU Rural Observatory, whose main objective is to further improve data collection and analysis on rural areas, but first results are expected by the end of 2022. This observatory is intended to increase the quantity and quality of available data as this is essential to understand the rural conditions to act on them properly.
Photo author: Santiago Sierra Durán (Salento, Colombia)
Rural areas are facing challenges such as ageing and depopulation. Heritage based regeneration plans can contribute to the sustainable development of these rural areas. This is a complex task, however, where a trade-off among the different regeneration plans and the limited available resources should be found and where computational methods can be useful to predict the best strategy.
One common approach when facing situations like this is through the analysis of some selected best practices or success stories (aka Role Models), and how innovation activities and cross-cutting themes successfully interacted in these Role Models. Then, these lessons learnt are adapted and replicated in other rural areas )aka replicators) for supporting the creation and implementation of heritage-led regeneration strategies.
In order to get quantifiable evidences, compara and appraise the effectiveness, impact and validity of the heritage-led regeneration actions, it is necessary to establish a robust monitoring systems based on a set of selected corss-thematic and multiscale Key Performance Indicators (KPIs) and evaluation procedures that ensure the production of a solid and reliable impact assessment of the strategies. Parameters obtained from role models and replicators baseline have been used to define an initial set of KPIs, which has been used for the first appraisal of the replicators baseline.
The methodology developed here allows to analysing an initial set of indicators as large as needed and, via several objective criteria, reduce the set of KPIs to a number that can be easily handled. But probably, the resulting set of KPIs will be diverse and not so easy to combine or compare, so group decision making techniques are useful to reach a trade-off among the experts´ opinions about how to combine the data from the indicators and get meaningful KPIs.
The impact of the strategies is assessed through KPIs in terms of Cultural and Natural Heritage according to the Communities Capital Framework (CCF). The KPIs intially considered for each replicator are re-tailored and further analysed by means of System Dynamics (SD), a suitable modelling technique for dealing with the nonlinear behaviour of complex systems over time suing sotcks, flows, internal feedback loops and time delays.
The RURITAGE project has identified 6 Systemic Innovation Areas (pilgrimage; sustainable local food production; migration; art & festivals; resilience; and integrated landscape management) which, integrated with cross-cutting themes, show case heritage potential as an engine for economic, social and environmental development of rural areas. CARTIF is in charge of developing the monitoring platfomr for assessing the impact of the action plans to regenerate the rural areas. Several dashboards have been designed focusing on KPIs values and their evolution4. RURITAGE has developed and set up a monitoring scheme to assess the performance pf the deployed regeneration action plans in six replicators. Performance monitoring is still ongoing and will last 2.5 years within project life.
1 RURITAGE, Rural regeneration through systemic heritage-led strategies, 2018. (https://www.ruritage.eu) Horizon 2020, Grant agreement No 776465.
2 Consejería de Cultura y Turismo, Plan PAHIS 2020 del Patrimonio Cultural de Castilla y León, Junta de Castilla y León. Consejería de Cultura y Turismo, 2015.
It is common knowledge that the moon goes through phases depending on its relative position between the earth and the sun. Thanks to that nights can be a showcase for looking to the starry skies or the perfect environment so that lycanthropes can take on vampires.
In science there are also phases, and the phase shifting, relative to the state in which matter is found. However, changes in this case have to do with temperature and heat and not with the states of the moon.
State transitions, have an important advantage and is that they are produced at constant temperature, allowing the matter to acquire and give up heat without changing temperature and thus reducing the impact over the environment that surrounds them. Are changes that, unlike the transformation suffered by David Naughton in “American Men in London” (film that achieve an Oscar in 1981 to the best makeup), aren´ t visible, but they are perceived.
The application of phase change materials, in particular those that have inside the homes usual transition temperatures between 18ºC-25ºC, can be used as recoveries in walls with which can reach a bigger comfort by stabilising the inside radiant temperature. It´ s not rare to found homes that because of a bad insulation are like thermic vampires, they remove us the heat, increasing the energy bill.
Inside the SUDO-SUDOKET project, which objective is the development of Key Enabling Technologies (KET) applied to innovative buildings, phase change materials dissolved in mortars have been studied to check its effect over the inside comfort conditions, as well as the effect over the climatization consume.
The results of the project had led to conclusions such as that a better stabilization of inside temperatures are reached if the radiant temperature is improved and, moreover, a reduction in the consume of climatization equipment, reaching energy saving, working as if it were a ring of garlic tied around the neck of our air-conditioning system.
The same as our favourite satellite goes from new to full, the enclosures of our homes will evolve to a future with a better control in superficial temperature and evenmore with adaptative enclosures that change of phase depending on the outside conditions.
Acknowledgments
The work has been done inside SUDOKET – mapping, consolidation and dissemination of Key Enabling Technologies (KETs) project for the construction sector at the SUDOE space, ref: SOE2/P1/E0677 that is co-financed by the Europeand Found of Regional Development (FEDER) through the INTERREG SUDOE programme.
Computer vision is a discipline that has made it possible to control different production processes in industry and other sectors for many years. Actions as common as the shopping process in a supermarket require vision techniques such as scanning barcodes.
Until a few years ago, many problems could not be solved in a simple way with classical vision techniques. Identifying people or objects located at different positions in images or classifying certain types of inhomogeneous industrial defects were highly complex tasks that often did not provide accurate results.
Advances in Artificial Intelligence (AI) have also accompanied the field of vision. While Alan Turing established the Turing test in 1950, where a person and a machine were placed behind a wall, and another person asked questions trying to discover who was the person and who was the machine, in computer vision through AI, systems capable of reproducing the behaviour of humans are sought.
One of the fields of AI is neural networks. Used for decades, it was not unitl 2012 that they began to play an important role in the field of vision. AlexNet1 , designed by Alex Krizhevsky, was one of the first networks to implement the 8-layer convolution filter design. Years earlier, a worldwide championship had been established where the strongest algorithms tried to correctly classify images from ImageNet2 , a database with 14 million images representing 1,000 different categories. While the best of the classical algorithms, using SIFT and Fisher vectors, achieved 50.9% accuracy in classifying ImageNet images, AlexNet brought the accuracy to 63.3%. This result was a milestone and represented the beginning of the exploration of Deep Learning (DL). Since 2012, the study of deep neural networks has deepened greatly, creating models with more than 200 layers of depth and taking ImageNet´ s classification accuracy to over 90% with the CoAtNet3 model. which integrates convolution layers with attention layers in an intelligent, deep wise way.
Turning to the relationship of modern computer vision models to AI, Dodge et. al (2017)4 found that modern neural networks classifying ImageNet images made fewer errors than humans themselves, showing that computer systems are capable of doing tasks better and much faster than people.
Among the most common problem solved by computer vision using AI are: image classification, object detection and segmentation, skeleton recognition (both human and object), one shot learning, re-identification, etc. Many of the problems are solved in two dimensions as well as in 3D.
Various vision problems solved by AI: Segmentation, classification, object detection
Classification simply tells us what an image corresponds to. So for example, a system could tell whether an image has a cat or a dog in it. Object detection allows us to identify several objects in an image and delimit the rectangle in which they have been found. For example, we could detect several dogs and cats. Segmentation allows us to identify the boundaries of the object, not just a rectangle. There are techniques that allow us to segment without knowing what is being segmented, and techniques that allow us to segment knowing the type of object we are segmenting, for example a cat.
Skeletal recognition allows a multitude of applications, ranging from security issues to the recognition of activities and their subsequent reproduction in a robot. In addition, there are techniques to obtain key points from images, such as points on a person´ s face, or techniques to obtain three-dimensional orientation from 2D images.
Industry segmentation using MaskRCNN5
One Shot Learning allows a model to classify images from a single known sample of the class. This technique, typically implemented with Siamese neural networks, avoids the need to obtain thousands of images of each class to train a model. In the same way, re-identification systems are able to re-identify a person or object from a single image.
The high computational cost of DL models led early on to the search for computational alternatives to CPUs, the main processors in computers. GPUs, or graphics processing units, which were originally developed to perform parallel computations for smoothly generating images for graphics applications or video games, proved to be perfectly suited to parallelising the training of neural networks. In neural network training there are two main stages, forward and back-propagation. During the forward process, images enter the network and pass through successive layers that apply different filters in order to extract salient features and reduce dimensionality. Finally, one or more layers are responsible for the actual classification, detection or segmentation. In backward propagation, the different parameters and weights used by the network are updated, in a process that goes from the output, comparing the obtained and expected output, to the input. The forward process can be parallelised by creating batches of images. Depending on the memory size of the GPUs, copies of the model are created that process all images in a batch in parallel. The larger the batch size we can process, the faster the training will be. This same mechanism is used during the inference process, a process that also allows parallelisation to be used. In recent years, some cloud providers have started to use Tensor Processing Units (TPUs), with certain advantages over GPUs. However, the cost of using these services is often high when performing massive processing.
Skeleton acquisition, activity recognition and reproduction on a Pepper robot6
CARTIF has significant deep neural network training systems, which allows us to solve problems of high computational complexity in a relatively short time. In addition, we have refined several training algorithms using the latest neural networks7 . We have also refined One Shot Learning systems using Siamese networks8. We also use state-of-the-art models in tasks such as object and human recognition, segmentation and detection, image classification, including industrial defects, and human-robot interaction systems using advanced vision algorithms.
1Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2012). Imagenet classification with deep convolutional neural networks. Advances in neural information processing systems, 25.
2Russakovsky, O., Deng, J., Su, H., Krause, J., Satheesh, S., Ma, S., … & Fei-Fei, L. (2015). Imagenet large scale visual recognition challenge. International journal of computer vision, 115(3), 211-252.
3Dai, Z., Liu, H., Le, Q., & Tan, M. (2021). Coatnet: Marrying convolution and attention for all data sizes. Advances in Neural Information Processing Systems, 34.
4 Dodge, S., & Karam, L. (2017, July). A study and comparison of human and deep learning recognition performance under visual distortions. In 2017 26th international conference on computer communication and networks (ICCCN) (pp. 1-7). IEEE.
5He, K., Gkioxari, G., Dollár, P., & Girshick, R. (2017). Mask r-cnn. In Proceedings of the IEEE international conference on computer vision (pp. 2961-2969).
6Domingo, J. D., Gómez-García-Bermejo, J., & Zalama, E. (2021). Visual recognition of gymnastic exercise sequences. Application to supervision and robot learning by demonstration. Robotics and Autonomous Systems, 143, 103830.
7Domingo, J. D., Aparicio, R. M., & Rodrigo, L. M. G. (2022). Cross Validation Voting for Improving CNN Classification in Grocery Products. IEEE Access.
8Duque Domingo, J., Medina Aparicio, R., & González Rodrigo, L. M. (2021). Improvement of One-Shot-Learning by Integrating a Convolutional Neural Network and an Image Descriptor into a Siamese Neural Network. Applied Sciences, 11(17), 7839.
