Improvement of the road maintenance through Aritificial Intelligence

Improvement of the road maintenance through Aritificial Intelligence

We all know that roads are necessary but normally we only remember them when they found them in bad conditions. We take it for granted that must always be available and in perfect condition, but this requries a great effort in terms of personnel, time and material resources. The spanish roads give support to the 86% of the inland transport of goods and to the 88% of the passenger transport. This high load of vehicles using the roads, together with the weather and environmental conditions cause a high level of wear with the consequent loss of properties of the road.

This cause to the users a series of severe inconveniences: the primary one is that it means a reduction in road safety, but also leads to a decrease in travel comfort, an increase of the fuel consumption of vehicles with the consequent increase of polluting gases emissions.

It is evident that the rehabilitation, preservation and maintenance of the road infrastructures is of fundamental importance, although we all know how annoying is founding roadworks. In Europe, in particular in Spain we have a good road grid, quite dense and good conected but certainly aged because of the decrease of the expenditure in preservation of the last years. It should be remembered that it requires a high level of investment in road maintenance; it is estimated, according to ACEX, that the annual maintenance cost of a motorway is of 80,000€, that of a conventional road of 38,000€ and that our country carries a preservation deficit of 8,000 million euros. This deficit, without going any further, it seems that will mean the approvement of tolls on motorways as of 2024. Therefore, these economic aspects and the need of a high level of service in roads demanded by the logistic and tourism sector, but especially the need of having safe roads, make the application of new technologies that can provide innovative solutions in road maintenance are in high demand.

The modern management of roads involves planning the maintenance actions to be carried out before the appearance of very serious or irreparable damage. This approach allows to undertake the interventions in the most adequate moment, causing as little inconvenience as possible and maintaining the fucntional capacity of the road and its economic value without allowing the network to be ruined and decapitalised. It is true that exist traditional solutions for the road preservation that are effective but it doesn´t make a optimus use of the available resources and it doesn´ t take in count the expected frime developments for planning the optimal time for action. To act effectively, is fundamental in first place to know the status of the road network as accurately and objectively as possible. This knowledge generally is obtained through road inspection equipments that make possible the evaluation and measurement of the corresponding parameters. In this way it achieves a large quantity of data related with the road status that it is necessary to manage and interpret to be able to prioritise the maintenance and preservation activities to be carried out. The problem that then arises is the processing of a massive quantity of information that makes impossible the manual evaluation.

One of the most difficulties, therefore, is the extraction of useful information of numerous data sources, For some type of data, exist software packages capable of extractinf global index that are useful for knowing in a general way the actual status of the road, but these tools often lack the capcity of predicting the road status evolution and its future degradation.

The artificial intelligence is becoming more and more present in a lot of areas of our environment and, often, without being conscious of it. The application of these artificial intelligence techniques can mean also a strong impact in the road maintenance because it allows the extraction of precise information of different data sources and identify relationships between them that otherwise could go unnoticed with the techniques applied until now. The processing and analysis, through the convolutional neural networks, of all the available data (data from the road auscultation equipment, climatological data, of traffic intensity…) allows obtaining unachievable data with the traditional methods. When training and adjusting those networks using massive quantity of data can be obtained, for example, highly reliable pavement degradation models that allow accurate estimation of the most appropriate maintenance actions.

road artificial intelligence

In this context, CARTIF and the company TPF actively collaborate in the development of these type of tools that can make a major breakthrough in improving road maintenance. Also there are other innitiatives that nowadays work in similar applications as Roadbotics (a spin-off of the Carneige Mellon University), the spanish company ASIMOB, Waterloo University in Canada, the finnish company Vaisala or the american company Blyncsy.

These tools will not eliminate the need of urgent repairs, as they can have many and varied origins, but it does have a significative impact on preventive and predictive interventions by making it possible to anticipate road deterioration and thus significantly reduce maintenance costs, reduce the time the road will be unavailable ad improve the degree of road comfort perceived by road users.

There are, finally, other interesting examples on how the artifical intelligence tools can help in the maintenance and improvement of the road safety, as for example the work of the MIT for predicting the road points in which it can occur traffic accidents and acting in consequence or the innitiative AI for Road Safety that use the artificial intelligence for reducing the number of road accidents.

In conclusion we can say that, thanks tot he help of these aritificial intelligence tools, in the next years we are going to have more safe and oeprative roads at the same time that we will notice that we found less works in our trips.

New technologies applied to security in confined spaces

New technologies applied to security in confined spaces

Ensuring the safety of workers inside confined spaces is a critical activity in the field of construction and maintenance because of the high risk involved in working in such environments. Perhaps it would be useful, first of all, to know what is meant by confined spaces. There are two main types of confined spaces: the so-called ‘open’ ones, which are those with an opening in their upper part and of such a depth that it makes their natural ventilation difficult (vehicle lubrication pits, wells, open tanks, tanks),…) and ‘closed’ ones with access openings (storage tanks, underground transformer rooms, tunnels, sewers, service galleries, ship holds, underground manholes, transport tanks, etc.). Workers entering these confined spaces are exposed too much greater risks than in other areas of construction or maintenance and it is therefore essential to apply extreme caution.

Each confined space has specific characteristics (type of construction, length, diameter, installations, etc.) and specific associated risks, which means that they require solutions that are highly geared to their specific safety needs.

The ‘conventional’ risks specific to confined spaces are mainly oxygen suffocation, inhalation poisoning of pollutants and fires and explosions. But new ’emerging’ risks from exposure to new building materials such as nanoparticles and ultrafine particles are also emerging. In addition, as research into new materials improves, there is also a better understanding of their potential negative effects on human health and how to prevent them.

The truth is that the training of workers and current safety regulations seek to anticipate risk situations before they occur in order to avoid them and thus prevent the appearance of accidents. But several problems arise: on the one hand, the regulations are not always strictly observed (whether due to workload, carelessness, fatigue, etc.) and on the other hand, there are always inevitable risks. In the case of carelessness, systems can be proposed to minimise this type of error and in the case of risks that cannot be avoided, systems can be proposed to detect them early and plan the corresponding action protocols.

It should be noted that risk situations do not usually appear suddenly and in most cases are detectable in time to avoid personal misfortunes. There are several problems: the detection of these risks is usually done with specific measurements using the portable equipment that the workers must carry, many times the workers are not controlled to access the premises with the corresponding protection equipmente and almost never a continuous monitoring of the indoor atmosphere is done.

In recent years, new technologies and equipment have been developed that can be applied to improve security in this type of environment and reduce the associated risks.

In this type of environment, an effective risk prevention system should be based on technological solutions capable of providing answers to safety aspects throughout the entire work cycle in confined spaces: Before entering the space itself, during all work inside the enclosure and when leaving the work space (whether it is at the end of normal work or by evacuation).

The latest confined space air quality monitoring systems are based on multisensorial technology that combine different detection systems to ensure the best possible conditions to avoid or reduce the risks present in the confined spaces.

Advanced data processing techniques (machine learning, data mining, predictive algorithms) are also being applied, enabling much more efficient and rapid information extraction.

In the same way, great advances have been made in access control and personnel tracking systems, allowing us to know the position of each worker and even his or her vital signs in order to detect almost immediately any problem that may arise.

Finally, it should be noted that the use of robots and autonomous vehicles (land and air) equipped with different types of sensorization are increasingly being used to determine the conditions of a site before it is accessed. This is especially useful in those where there may have been an incident: power failure, collapse, fire,… or simply because environmental conditions are suspected to have changed and the reason is unknown.

CARTIF has been working on these issues for many years now, both in safety projects in critical construction environments (PRECOIL, SORTI) and in specific systems for tunnels and underground works (PREFEX, INFIT, SITEER).

In short, the development and implementation of new specific technologies can help to save lives in such a critical environment as confined spaces.

New applications of Deep Learning

New applications of Deep Learning

A little more than a year ago, in another post of this blog, our colleague Sergio Saludes already commented what is deep learning and detailed several of its applications (such as the victory of a machine based on these networks over the world champion of Go, considered the most complex game in the world).

Well, in these 16 months (a whole world in this topic) there has been a great progress in terms of the number of applications and the quality of the results obtained.

Considering, for example, the field of medicine, it has to be said that diagnostic tools based on deep learning are increasingly used, achieving in some cases higher success rates than human specialists. In specific specialties such as radiology, these tools are proving to be a major revolution and in related industries such as pharmaceuticals have also been successfully applied.

In sectors as varied as industrial safety, they have recently been used to detect cracks in nuclear reactors, and have also begun to be used in the world of finance, energy consumption prediction and in other fields such as meteorology and the study of sea waves.

Autonomous vehicle driving projects, so in vogue these days, mainly use tools based on deep learning to calculate many of the decisions to be made in each case. Regarding this issue, there is some concern about how these systems will decide what actions to take, especially when human lives are at stake and there is already a MIT webpage where the general public can collaborate in creating an “ethics” of the autonomous car. Actually, these devices can only decide what has previously been programmed (or trained) and there is certainly a long way to go before the machines can decide for themselves (in the conventional sense of “decide”, although this would lead to a much more complex debate on other issues such as singularity).

Regarding the Go program discussed above (which beat the world champion by 4 to 1), a new version (Alpha Go Zero) has been developed that has beaten by 100 to 0 to that previous version simply knowing the rules of the game and training against itself.

In other areas such as language translation, speech comprehension and voice synthesis have also advanced very noticeably and the use of personal assistants on the mobile phone is beginning to become widespread (if we overcome the natural rejection or embarrassment of “talking” with a machine).

CARTIF is also working on deep learning systems for some time now and different types of solutions have been developed, such as the classification of architectural heritage images within the European INCEPTION project.

All these computer developments are associated with a high computational cost, especially in relation to the necessary training of the neural networks used. In this respect, progress is being made on the two fronts involved: much faster and more powerful hardware and more evolved and optimized algorithms.

It seems that deep learning is the holy grail of artificial intelligence in view of the advances made in this field. This may not be the case and we are simply looking at one more new tool, but theres is no doubt that is an extremely powerful and versatile tool that will give rise to new and promising developments in many applications related to artificial intelligence.

And of course there are many voices that warn of the potential dangers of this type of intelligent systems. The truth is that it never hurts to prevent the potential risks of any technology, although, as Alan Winfield says, it’s not just artificial intelligence that should be feared, but artificial stupidity. Since, as always happens in these cases, the danger of any technology is in the misuse that can be given and not in the technology itself. Faced with this challenge, what we must do is promote mechanisms that regulate any unethical use of these new technologies.

We are really only facing the beginning of another golden era of artificial intelligence, as there have been several before, although this time it does seem to be the definitive one. We don’t know where this stage will take us, but trusting that we will be able to take advantage of the possibilities offered to us, we must be optimistic.

Geolocation systems are reaching indoors

Geolocation systems are reaching indoors

With global positioning systems, a phenomenon similar to what happened with mobile phones has occurred: in a few years we have gone from non-existence to consider it essential. The truth is that, in fact, geolocation is one of those technologies that has led to the development of many applications and in many areas is not conceived to work without the use of commonly called GPS.

These types of positioning systems are based on receiving the signal from three or more satellites and using trilateration: position is obtained in absolute coordinates (usually WGS84) by determining the distance to each satellite.

satelite-geolocalizacion

Global positioning systems based on satellites have their origin in the US system TRANSIT  in the 60s. With this system you could get fix the position once an hour (at best) with an accuracy of about 400 meters. This system was followed by the Timation system and in 1973 the Navstar project began (both from USA). The first satellite of this project was launched in February 1978 and full operational capability was declared in April 1995. This Navstar-GPS system is the origin of the GPS generic name we usually apply to all global navigation systems. In 1982 the former Soviet Union launched the first satellite of a similar system called GLONASS that became operational in 1996. Meanwhile, the People’s Republic of China in 2000 launched the first satellite of BeiDou navigation system, which is scheduled to be fully operational in 2020. Finally, in 2003, it began the development of the positioning system of the European Union called Galileo, with a first launch in 2011. Currently there are 12 satellites in active (and 2 in tests) and the simultaneous launching of four more is scheduled on 17 November 2016. This way, 18 satellites will be in orbit and initial service of Galileo positioning system could begin in late 2016. It is expected to be fully operational in 2020. It must be said that there are also other systems, complementary to those already mentioned, in India and Japan in a local range.

As you can see, the global positioning systems are fully extended and are widely used both military and commercial level (transport of people and goods, precision agriculture, surveying, environmental studies, rescue operations …) and on a personal level (almost everyone has a mobile phone with GPS available, although their battery always run out at the worst moment).

Regarding the precision obtained with current geolocation equipment, it is about a few meters (and even better with the Galileo system) and can reach centimetre accuracy using multifrequency devices and applying differential corrections.

geolocation-system

One of the problems of these systems is that not work properly indoors since the satellite signal cannot be received well inside buildings (although there are highly sensitive equipment that reduce this problem and other devices called pseudolites, acting simulating the GPS signal indoors). And of course it’s not enough to know our exact position outdoors but now comes the need to also be located inside large buildings and infrastructure (airports, office buildings, shopping centres, …).

So indoor positioning systems (IPS) have appeared allowing location inside enclosed spaces. Unlike global positioning systems, in this case there are many different technologies that are usually not compatible with each other making it difficult to dissemination and adoption by the general public. There are already very reliable and accurate solutions in enterprise environments but these developments are specific and not easily transposable to a generic use of locating people indoors. In this type of professional context, CARTIF has done several projects indoor positioning for autonomous movement of goods and service robotics. There is not a standard indoor positioning system but there are many technologies competing for a prominent place.

The technologies used can be differentiated on the need or not of a communications infrastructure. Those who no need existing infrastructure are often based on the use of commonly available sensors in a smartphone: variations in the magnetic field inside the building that are detected by the magnetometer, measuring the movements by using accelerometers or identifying certain feature elements (such as QR codes) using the camera. In all these cases the accuracy achieved is not very high but may be useful in certain applications as simple guidance in a large building.

Indoor positioning systems using communications infrastructure exploit almost all available technologies of this kind for the location: WiFi, Bluetooth, RFID, infrared, NFC, ZigBee, Ultra Wideband, visible light, phone masts (2G / 3G / 4G), ultrasound, …

geolocalizacion-smartphone

With these systems, the position is usually determined by triangulation, calculating the distance to the fixed reference devices (using the intensity of the received signal, coded signals or by direct measurement of this distance). Thus you can reach greater precision than in the three previous cases. There are also new developments that combine several of the above technologies in order to improve the accuracy and availability of positioning.

Although, as has been said there is no standard, the use of systems based on Bluetooth low energy are spreading (BLE nodes). Examples of such systems are the Eddystone (from Google) and iBeacons (Apple).

Logically, as in the case of outdoor positioning the corresponding environment map is required to allow navigation. There are other systems, called SLAM, which generate environment maps (which may be known or not) as they move, widely used in robots and autonomous vehicles. A recent example is the Tango project (from Google once again) that generates 3D models of the environment just using mobile devices (smartphones or tablets).

As we have seen, we are closer to be located anywhere, which can be very useful but also can make us overly dependent on these systems while the usual privacy issues concerning positioning systems are increased. So although thanks to these advances the sense of orientation is less necessary, we must always keep common sense.

The future of construction is printed in 3D

The future of construction is printed in 3D

3D printing is here to stay. When a new technology is so widespread that no longer catches the attention it is that its implementation is complete. More and more people have a plastic 3D printer at home and many of us know someone who has bought one or it has been built by pieces. It was only a matter of time before this technology would give the jump to other sectors. Although the construction sector usually adopts this type of technological developments rather late, in this case there are already several projects trying to bring the additive manufacturing (as is also known 3D printing) to construction.

What is wanted, among other things, it is to face the new architectural designs that are increasingly complex, industrialize certain construction processes which, today, are almost artisanal and improve sustainability using recycled materials for printing.

Such systems pose major challenges such as the development of new building materials that allow their proper implementation. Usually, the addition of other materials or compounds that improve the properties (or achieve the desired properties) in setting times, strength and insulation is used.

One of the first projects in relation to additive manufacturing in construction is called “Contour Crafting“, led by Dr. Behrokh Khoshnevis of the University of Southern California. And now there are many research centres and universities focused on these issues as AMRG University of Loughborough considered a world reference or IAAC in Spain.

They have also appeared commercial developments such as the case of a Chinese company that manufactured homes, offices and entire buildings using these techniques. The specific case of this company seems to respond to marketing strategies (which seems to be taking effect) because a good position in these technologies can open important markets.

In any case, there are many interesting initiatives such as WASP, an Italian project for sustainable buildings in disadvantaged areas, the construction of a steel bridge in Amsterdam, or NASA contest for construction of buildings on the moon or Mars using these techniques, the winner of which proposed the use of ice as raw material.

In the light of these developments it is easy to see that the additive manufacturing construction offers some advantages hard to match with other methods such as complexity in designs that can be obtained, the accuracy and repeatability of certain construction procedures. It is undeniable that industrialization is increasingly integrated into many building processes and 3D printing sure to have your niche in the construction sector.

As always with new technologies, certain optimistic sectors are saying that the additive manufacturing will be the majority system used in all industries but certainly there are currently no universal manufacturing technologies (beyond certain methods such as mass production). The current manufacturing processes are highly specialized and uses the most appropriate technologies in each case it seems complicated than a single technology is able to replace almost all existing. Therefore, and being realistic, we must find the most suitable application field for 3D printing in construction.

In this regard, CARTIF participates in a major national research project related to 3D printing in construction. This project focuses on the application of 3D printing technologies in construction in those areas where it is considered that can be especially useful: the manufacture of prefabricated modules and rehabilitation of facades. It does not seek a universal technology to serve in all areas of construction, but to reach the market with a product that offers a viable alternative to other existing technologies (i.e. realistic and sustainable applications). And without forgetting that all progress made in this field (whether by R & D or marketing strategies) will impact in the future, for the benefit of the whole society because what it is pursued, is to build better, faster, cheaper and in a more sustainable way.