Big Data, Blockchain and Artificial Intelligence
Big data plays a central role in processing the enormous volume of information that today makes up an industry’s information system. Indeed IoT, MES platforms, optimization algorithms, ERP today make up a gargantuan volume of data . Conventional data storage technologies are now overtaken by these volumes.This is why appropriate technologies have been created to store such volumes, such as object databases or server infrastructures adapted to distributed processing. Several reasons are at the origin of these increasingly large quantities of data, but the development of the Internet and data exchange infrastructures are certainly at the origin of the creation of this need in companies.
So there are now solutions that can meet this need.
However, this is useless if there is no system to process, analyse and extract this data.
This technology has led to the creation of a new profession: Data scientist . A data scientist aims to create ever more efficient algorithms to extract relevant and usable data from Big Data platforms .Data is central to the concept of smart manufacturing and solutions that have proven their effectiveness now exist to facilitate processing. However, this volume of data and the different sources bring a new technical and structural problem: security. Blockchain technologies could well bring their stones to the building.
While the volume of data and the number of flows reported by production lines, machines and various sensors or IoT are only increasing, industrial applications require a high degree of trust and protection. Indeed, proof of the existence of the data source (sensor, machine, etc.) or of the consistency and veracity of the data reported becomes very important.
How can we be sure that the data we receive from a sensor, for example, is real, unmodified on its way and coming from a single interlocutor, verified, valid and expected?
It is in this context that the blockchain brings real value to the data.
Blockchain is a “cryptography-based computer protocol” that secures digitally transferred data from a single data source to a single recipient.The blockchain brings confidence in the veracity of the source and the data through several aspects:”Everyone” (meaning all “subscribers” of the blockchain) has a complete log of the exchanges and can control that the transfer has been madeNo one can refute the legitimacy of the transfer
Finally, the blockchain brings confidence in the data by not centralising authorizations and authentication. All of these processes are “distributed” over a network, which reduces the risk of errors or corruption. Thus, this new technology makes it possible to solve the problems of trust in relation to data.
Today’s industrial world requires ever more information and data exchange. This boom in data calls for a revolution in the processing and use of data. It is to this problem that artificial intelligence tries to respond within the industry .It is no longer a question of summarising data as we know it today via dashboards and graphs. It is necessary to exploit all this data to deduce trends and make predictions on the behaviour of the data over time: this is what artificial intelligence offers today in the industry.The platforms offered by professionals are still not very intelligent, far from the experiments offered by Google , or the famous Watson from IBM. But they show the path that could soon be taken by applications that integrate real pieces of AI.
However, some solutions are beginning to emerge. They are designed based on deep learning . The methods associated with deep learning or deep learning, allow automatic learning with a high level of abstraction linked to the observation of data. This observation helps to give some meaning to the data.This data can for example allow artificial intelligence to optimise a production chain , or even to predict the risks of slowdown or failure. Artificial intelligence now makes sense thanks to the Internet of Things, which makes it possible to collect large amounts of data.
However, there are still some challenges for AI, the biggest being automated decision making. It is here that artificial intelligence plays an increasingly important role and many players are offering solutions to introduce it to the heart of the industry. While this technology and the resulting applications are still in their infancy, they nevertheless provide fast and effective solutions in “guided” decision-making .By giving meaning to data, artificial intelligence will no longer only be an actor in analysis but also becomes an actor in automated decision-making. It is then that these intelligent functions will be able to intervene effectively on systems such as:
The difficulty of this type of implementation is to find the right balance between intelligent and automated decision-making and that of the human. This is quite a challenge that still needs to be tackled in smart manufacturing.
Today machines are always more complex in order to meet the performance, efficiency and quality needs of production lines. The complexity of maintaining these machines has therefore increased exponentially. Augmented reality, as shown in the example of the photo above, is a facilitator and makes it possible to inform and train in a didactic and visual way the operators at the heart of production. Each time the touchpad is moved, the operator visualises the machine he is observing in real time and