WO2025068377A1 - Method for training a neural network intended for controlling a system for sorting objects - Google Patents

Abstract

The invention relates to a method (100) for processing and characterising objects by means of a neural network, the method comprising generating (150) items of neural network digital training data (T.DAT) representative of simulated objects to be processed (S.IT), which are generated by a digital twin of the object processing system and are associated with annotations (ANN) that indicate at least one characteristic of the simulated objects (S.IT); training (160) the neural network by means of the items of digital training data (T.DAT); acquiring (170) digital images of a real scene comprising at least one real object to be processed; providing (180) items of data (R.DAT) generated from the digital images of the real scene to the neural network, which in return outputs an item of data representative of a characteristic (P.G.O) of the real object; and sending (190) a control signal to an object processing device.

Description

Training method for a neural network intended for controlling an object sorting system TECHNICAL FIELD OF THE INVENTION

The invention relates to a method for processing objects, such as a sorting method, the method including training a neural network intended to control an object processing system, said neural network being trained using images comprising objects to be processed each associated with a characteristic used as a processing criterion.

TECHNOLOGICAL BACKGROUND

Object handling systems such as baggage, parcel, postal or other object sorting systems use conveyors and/or gripper robots as handling means. These handling means are controlled using various types of sensors such as photoelectric cells, scales, force sensors, barcode readers or digital cameras connected to electronic and/or computer systems.

Among the methods for controlling these sorting systems, the use of artificial neural networks is considered a very effective solution that can have high reliability, with a very low error rate. One example is the control of gripping robots using images captured by digital cameras. In this case, the neural network has the function of characterizing, possibly by identification, objects to be sorted found in the images captured by the cameras in order to control in return gripping robots whose function is to move these objects. Thus, so-called "deep learning" methods are developed to process complex stacks of items to be sorted, as described in the article "Visual Sorting of Express Parcels Based on Multi-Task Deep Learning" by Song Han et al., Sensors 2020, 20, 6785; doi:10.3390/s20236785.

Neural networks are trained on sets of manually labeled images to adjust their internal parameters, such as the synaptic weights of each of the neurons constituting the network. Image labeling involves clipping the objects of interest by creating a mask excluding parts of the image other than these objects, and then annotating each of the objects. For quality training, the training data should contain as many images as possible, with 100,000 images being considered a minimum. However, manually processing and labeling these images is time-consuming and prone to human error. It is therefore difficult and expensive to produce large sets of labeled data for training neural networks.

In view of these difficulties, patent document WO 2018/170512 A1 discloses a solution based on a sequence of images, in which an object is manually labeled in a first image of the sequence, and then this object is automatically tracked and labeled in subsequent images of the sequence, thus providing a large number of training images for limited human effort. It is also proposed to carry out automatic annotations of images, as for example in patent document EP 3 885 998 A1.

However, a satisfactory solution for sorting applications has yet to be implemented.

An object of the invention is a method of object processing including the training of a neural network based on the use of automatically produced annotated images.

To achieve this aim, a first aspect of the invention is a method for processing objects in an object processing installation comprising a computer configured to characterize, by means of a neural network, objects to be processed visible in digital images of objects to be processed located in the object processing installation, the method comprising: a step of generating digital training data representative of objects to be processed simulated within the object processing installation, these data being intended for training the neural network, generated by means of a digital twin of the object processing system, and associated with annotations indicating at least one characteristic of the simulated objects to be processed, the digital twin integrating a dynamic engine for simulating physical interactions between at least one object to be processed and at least one simulated element of the processing installation, so as to simulate movements of objects to be processed and interactions of these objects to be processed with the at least one simulated element of the processing installation; a step of training the neural network comprising providing digital training data as input to the neural network and during which the parameters of the neural network are adjusted on the basis of the digital training data; a step of acquiring digital images of a real scene of the object processing installation comprising at least one real object to be processed; a step comprising providing data generated from the digital images of the real scene as input to the neural network, the neural network providing in return data representative of a characteristic of the real object to be processed; and a step comprising sending a control signal to an object processing device on the basis of the data representative of the characteristic of the at least one real object to be processed.

This method benefits from the automatic, simple and economical production of a large number of annotated images adapted to the purpose of the method, which is the embodiment detailed below consisting of sorting articles arriving on an article sorting line. Thus, the training of the neural network, based on a large number of annotated images without errors and generated in such a way as to correspond to the situation for which the neural network is used, is of high quality, and the use of this neural network after its training implies a high level of reliability of the sorting process to which it is applied.

According to additional non-limiting characteristics of the first aspect of the invention, considered individually or in any technically feasible combination:

- the simulations of movements of objects to be processed and the simulations of interactions of the at least one object to be processed with the at least one element of the processing installation may comprise a simulation of a physical interaction between the at least one object to be processed and at least one conveyor of the installation, so as to simulate a configuration of the at least one object to be processed in a work area of the sorting installation;

- the at least one simulated element may be a ginning device;

- the movements of simulated objects to be processed and the interactions of the at least one simulated object to be processed with the at least one simulated element may include a simulation of unloading the objects to be processed in an unloading zone;

- the at least one characteristic of the simulated objects to be processed can be chosen from the position, geometry and orientation of the simulated objects to be processed;

- the digital training data may comprise at least one digital image of a simulated scene representing a working area of the object processing facility;

- the at least one digital image can be generated so as to be substantially as if it were obtained from digital images acquired by a digital camera installed above the working area of the actual processing installation;

- the digital training data may further include a depth map of the simulated scene;

- the depth map may be generated so as to be substantially as if it were obtained from digital images acquired by a digital camera installed above the working area of the actual processing facility;

- the digital twin integrates a dynamic engine for simulating physical interactions between simulated elements of the treatment installation, so as to reproduce movements and shocks between simulated objects to be treated and simulated elements of the treatment installation;

- the method may further comprise a step of projecting textures from digital images of real objects onto surfaces of the simulated objects to be processed;

- the digital twin of the object processing system can be configured so as to generate digital images of homogeneous batches of objects to be processed comprising more than 40%, preferably more than 50%, preferably more than 80%, preferably more than 90% or 100% of objects belonging to the same type of objects to be processed; and

- the type of objects can be defined by a texture mapped onto surfaces of simulated objects.

The invention extends to a treatment installation comprising a computer configured to implement the method according to the invention.

In the treatment facility, (i) the computer can be configured to control the manipulator arm and (ii) the computer and the digital twin can be interfaced in such a way that at least one simulated element of the facility is controlled by the computer.

BRIEF DESCRIPTION OF THE FIGURES

Other characteristics and advantages of the invention will emerge from the detailed description of the invention which follows with reference to the appended figures in which:

There represents an overall view of an installation in which the system according to the invention is integrated;

There illustrates a mode of operation of the system according to the invention, using a top view;

There illustrates a method according to the invention, which can be implemented by the system of figures 1 and 2; and

There illustrates an example of part of the installation of the .

DETAILED DESCRIPTION OF THE INVENTION

Figures 1 to 3 illustrate a mode of implementation of a method for training a neural network, specialized in the analysis of digital images in order to identify items to be sorted within an item sorting line.

A context in which the invention can be applied is illustrated using the example of Figures 1 and 2, which represent a schematic portion of a FAC installation for processing articles comprising an SL sorting line equipped with a SYS system for controlling the sorting line. The straight arrows indicate the paths followed by articles to be sorted and being sorted within the installation.

There illustrates a sorting installation FAC, comprising a working zone WZ in which IT articles to be sorted brought by a truck Trck are poured in bulk into an unloading zone DZ, then brought to a working zone WZ by means of an article de-separating device DE. At the working zone WZ, a robotic manipulator arm MA as an individual article handling device is configured to de-separate the IT articles and place them in an orderly manner on a conveyor belt CB leading to an inserter INS, a circular conveyor belt CCB receiving the articles from the inserter INS and sorting outlets Out (three outlets shown here) to which articles are directed according to their final destination. This may, for example, involve sorting postal parcels and dividing them into delivery rounds, each corresponding to one of the sorting outlets.

There illustrates the work area WZ. The manipulator arm MA is equipped with a gripping head GH, and the conveyor belt CB moves the articles It unloaded from the bulk truck in the direction Dir represented by an arrow, to the inserter INS. The gripping head GH can be equipped with suction cups SC connected to a pumping system, each suction cup can be controlled individually so as to apply a vacuum to the surface of an article to be gripped. The articles arrive in bulk from the right of the figure, the manipulator arm being responsible for unpicking them, that is to say, gripping them one by one to place them in order on the conveyor belt CB towards the left of the figure. A system SYS for controlling the sorting line comprises a digital camera CAM and a digital computer CALC comprising a computer memory operatively connected to the digital camera CAM and the manipulator arm MA. The MA manipulator arm has the function of individually manipulating IT articles located in the work area to move them from this area to the CB conveyor belt.

We will focus on the process of identifying items in the work area for picking by the manipulator arm.

At this level, the articles are handled one by one in order to order them for subsequent sorting operations, in this case the insertion of the articles at regular intervals on the circular conveyor belt. This individual handling of the articles must be done as quickly as possible, while ensuring good quality of capture of the articles by the robotic arm. To this end, the system must provide the robotic arm control system with the characteristics, positions, orientations, and geometries of each of the articles. These characteristics are obtained from digital images of the work area, processed by the CALC computer using a NN neural network trained using annotated images.

Conventionally, these images are annotated manually, or at best, with semi-automated annotation systems, as mentioned in the Technological Background section. The solution proposed here is based on the use of a digital twin of the sorting facility to computationally generate digital images and depth maps representative of scenes encompassing the work area.

The use of a digital twin is particularly relevant in a situation where it is difficult to carry out manual annotation of training images. When it comes to associating a binary content annotation with images, such as “presence or absence of a yes/no address block”, then a human operator can be efficient in terms of time, reliability and cost to carry out the annotation. On the other hand, when the required annotation is for example to determine and indicate a position, geometry or dimensions of items present in an image, then the time, reliability and cost of a human operator are very unfavorable. In such situations, it is particularly interesting to implement an automatically annotated training image generation, which will be faster, more reliable and/or more economical than if it were carried out by a human operator.

There represents a diagram of the method 100 for handling articles according to the invention, including the generation of annotated images for training a neural network, the training of this network, then its use in controlling a robotic arm for picking articles to be sorted.

At a step 110A, the DAT data from the FAC article processing installation are collected, which includes the geometric and functional characteristics of the equipment used (sorters, conveyors, device for sorting or diverting articles, etc.) and the installation of this equipment, the interactions between these different elements.

In a step 120, based on the collected DAT data, a digital twin NUM.TWIN of the sorting installation is developed, stored in a computer memory of the CALC computer. This digital twin integrates a dynamic engine configured to reproduce the physical interactions of the simulated elements of the sorting installation (i) with each other and (ii) with the articles that this installation processes, as well as the physical interactions of the objects to be sorted (i) with each other and (ii) with the simulated elements of the sorting installation. As already mentioned, the simulated elements of the sorting installation are, for example, conveyors, devices for sorting or diverting the objects to be sorted, sorting subsystems forming part of the installation, or quite simply physical elements such as plates supporting the articles or barriers blocking their passage. In this way, the path, movements, shocks and more generally the interactions of the simulated objects between themselves and with the elements constituting their environment, and specifically the elements of the installation, are faithfully reproduced by the digital twin.

The dynamic engine can consist of a software grouping configured to reproduce the laws of dynamic mechanics. In combination with an automation reproducing the control system of the installation elements, the digital twin makes it possible to generate realistic configurations corresponding to actual processing of articles by the installation, in particular at the level of the WZ work zone. The dynamic engine allows for high realism and a wide variety of configurations, guaranteeing effective training of the neural network subsequently.

In a step 130, the digital twin generates simulated scenes of the sorting installation comprising in particular configurations of simulated articles to be sorted S.IT in the work zone WZ. The simulated scenes result from a physically realistic simulation of the interactions between the articles and the devices of the processing installation FAC, from the entry point of the articles into the simulated installation to the work zone WZ where they will be individually gripped by a manipulator arm MA. This exhaustive simulation of the processing of the articles within the installation FAC makes it possible to generate scenes extremely close to those which would be obtained within the real installation, as explained using the example of the .

There illustrates an example of a sorting installation such as that of the , for which the digital twin simulation takes into account: the bulk dumping of IT articles from a transport truck Trck onto a dumping zone DZ, and the transport of the dumped articles by means of a device DE for ginning the articles from the dumping zone DZ to the working zone WZ. The ginning device DE is, in this example, made up of three successive conveyors C1, C2 and C3 each equipped with drive motors M1, M2 and M3, respectively. The conveying speeds of the articles by the conveyors are controlled individually by driving the motors. For example, driving motor M2 allows the speed of conveyor C2 to be controlled. Each of the conveyors can have an inclination. This inclination and the driving of the motors of the device DE allow for progressive ginning (separation of the objects from each other) of the articles conveyed from the zone DZ to the zone WZ.

One objective of seeding is to facilitate the individual processing of articles in the working area WZ. In order to optimize the handling of articles by the arm MA, it is possible to play on the inclinations and the control of the motors, the variations of which lead to different configurations of the articles in the working area, facilitating more or less the work of the manipulator arm MA. This essentially concerns the spatial and geometric aspects of the configurations that the articles take in the WZ zone. More generally, these configurations can be characterized by one or more characteristics such as positions, orientations, and geometries of each of the articles in a given configuration, as well as by textures applied to volumes representing each of the objects, textures which can include logos, patterns, address blocks, or other visual elements distinguishing an article or a type of article from another article or another type of article. Of course, the generation of articles of the simulated scenes can be controlled according to conventional means such as simulation parameters, so as to reproduce the situations envisaged by the practitioner with regard to, for example, the shapes, dimensions or even textures of the articles, the density, the number or the distribution of other different types of these.

However, the advantage of exhaustive simulation of the processing of articles lies in the possibility it gives to obtain configurations of objects which are representative, on the one hand of the nature and arrangement of the elements forming the sorting chain, and on the other hand of their settings and their control algorithms. In the present example, the speeds of the motors M1, M2 and M3 of the conveyors C1, C2 and C3 can be adjusted, which will condition the quality of the ginning and therefore the configuration of the articles at the level of the zone WZ, at the exit of the ginning device DE. For the purposes of simulation, the working zone WZ and the discharge zone DZ can be considered as rigid horizontal plates with which the objects to be processed physically interact.

Furthermore, the digital twin can be interfaced directly with the SYS control system of the real FAC installation, which is simulated by the digital twin, in such a way that the elements of the simulated FAC installation are also controlled by the SYS system. This can be referred to as virtual commissioning of the FAC installation for processing articles. In this way, and taking into account the lessons learned in the previous paragraph, the scenes generated by simulation are brought even closer to the scenes that would actually be obtained in the real installation, the devices in the simulation being controlled in the same way as the real devices.

Generally speaking, it will be advantageous to generate a scene by simulating the interactions of the objects to be processed with the sorting facility equipment from their respective introductions into the facility to the work area. This approach makes it possible to obtain simulated scenes as close as possible to the scenes that would be obtained under the same conditions (characteristics of the objects, the equipment, the control algorithms) in the real system.

In a step 110B, digital images R.IMG of real articles representative of the articles intended to be sorted by the FAC installation are collected and stored on a computer storage medium. This operation may, for example, consist of taking digital photos of each of the surfaces of postal articles or parcel packaging such as those which are to be sorted, by sampling in the sorting installation of already sorted articles or at article packaging suppliers.

At a step 140, the CALC computer projects textures from the digital images R.IMG onto the surfaces of the simulated articles S.IT, thus subsequently making it possible to obtain photorealistic viewpoints of the simulated scene. At a step 150, the CALC computer generates a digital image and a depth map of the scene. This digital image and this depth map constitute annotated T.DAT training data, the annotations indicating the respective characteristics of the articles in the scene: position, geometry, orientation, for example. The digital image can be generated so as to include masks taking into account the occlusion of each of the articles.

The digital image and the depth map of the scene are preferably generated in such a way as to be substantially as if they were obtained from digital images acquired by the digital camera CAM in the real sorting facility. This means that the viewing angle of the simulated scene chosen to generate the digital image and the depth map corresponds to the viewing angle from which the digital camera CAM observes the real scene. The objective is to train the neural network with data as close as possible to the data obtained from real shots that it will have to process following its training period. The training can be done without the depth map, which is only optional, but can improve the quality of the training.

An important advantage of this training data is that, being entirely automatically generated by calculation, the characteristics of the items in the scene, generated during the simulation of the scene, are perfectly known and the annotations, also automatic, are immediately available and necessarily accurate, which gives this process a decisive advantage over conventional processes for producing annotated images.

An additional advantage of the method described in this paper is that the operator can choose to simulate expected scenes, in order to anticipate a potentially problematic situation.

An example is the arrival of a new type of packaging for items to be sorted, for example through the arrival of a new economic actor producing items to be sorted with a new design, such as a set of particular patterns on the packaging. In this case, this known design can be integrated in advance into the training data in order to prepare the neural network for sorting this new type of packaging.

Another example is the massive arrival of a type of items to be sorted, thus forming homogeneous batches of items to be sorted. An item type can be defined by a shape (cylindrical, parallelepiped, hexagonal), a texture (pattern that will be applied to the volume of a simulated item), a logo, a specific color, etc., or a combination of any number of these characteristics. An item type can include subtypes: in the case of a type defined by a parallelepiped shape, specific ratios between a length, a width, and a depth can form a subtype. Items of the same type can, for example, come from the same economic actor and have the same visual identity, characterized by the type to which the items belong, such as a texture of the packaging used by this economic entity.

Outside of this context of the massive arrival of articles of the same type, a usual situation could be that of heterogeneous batches of articles to be sorted, that is to say batches made up of articles of different types. In this usual situation, the articles can then be distinguished relatively easily from each other. On the other hand, in the case of the arrival of a homogeneous batch of articles, that is to say a batch of articles resembling each other, in particular because they belong to the same type of article, these articles can be difficult to distinguish from each other in the images used for sorting. Such a situation can be anticipated by generating training images comprising homogeneous batches of articles in order to train the neural network to manage this specific situation. By "images comprising homogeneous batches of articles" we mean images generated by the digital twin configured to generate images with articles of the same type. The operator can thus choose to generate digital training images of homogeneous batches of articles comprising more than 40%, more than 50%, more than 80%, more than 90% or 100% of articles belonging to the same type, according to conventional methods of adjusting the simulation parameters. Trained on the basis of these images, the neural network will be better prepared to process real images.

Steps 130 to 150 are repeated so as to obtain a number of images considered sufficient by the system operator for training the neural network.

At a step 160 of training the neural network, the annotated training data are provided by the CALC calculator as input to the neural network NN during its training phase. During this training phase, the internal parameters P _i of the network, which characterize its synaptic weights and therefore its response to data provided as input, converge towards values allowing a reliability deemed sufficient for the processing of these data, reliability estimated for example by a success rate of the network in processing the data and which must exceed a given threshold to be considered sufficient. This rate reflects the level of adequacy of the network's responses to the data provided compared to the expected responses, corresponding to the annotations.

During the completed training phase, the internal parameters of the neural network were adjusted to provide appropriate responses to the data provided as input to the network, and it can be used during the real item sorting process.

In a step 170, the digital camera CAM acquires digital images of a real scene encompassing the working area WZ in the form of a color digital image and two images intended to form a depth map of the scene. The computer CALC generates the depth map.

At a step 180, the resulting data R.DAT (unannotated) from step 170 comprising the color image and the depth map of a real scene are provided to the neural network NN. The neural network, trained for this task, then identifies the positions, geometries and orientations P.G.O of the articles in the scene. Generally speaking, the neural network could be designed and trained to identify any other characteristic of the article: a destination or origin address or an identification in a textual form or a digital code such as a bar code or a QR code, a logo, a category such as a type of packaging (polybag, cardboard, envelope, etc.). Furthermore, this principle can be applied to any kind of object such as an article, a parcel, a box, a letter, a package or natural product or a manufactured product.

At a step 190, on the basis of the positions, geometries and orientations P.G.O of the articles identified by the neural network, the computer CALC of the system SYS for controlling the sorting line sends a control signal to the robotic manipulator arm MA so as to unpick the articles by handling them individually according to at least one of their respective positions, geometries and orientations, according to conventional methods.

Steps 170 to 190 are repeated so as to process all the items appearing in the work area WZ.

In the implementation described above, a CALC calculator is used for all phases of generating training data, training the neural network, and during the item sorting process. It is not necessary for the same calculator to be used for these different tasks: each could be performed by a dedicated calculator, and the neural network could be transmitted from one to the other of these calculators.

The figures in this document are not necessarily to scale. Some features and components may be shown exaggerated in relation to other components or in a somewhat schematic form, and some details of conventional items may not be shown in the interest of clarity and conciseness.

Of course, the invention is not limited to the method of implementation described and variant embodiments can be made without departing from the scope of the invention as defined by the claims.

Claims (15)

Method (100) for processing objects in an object processing installation (FAC) comprising a computer configured to characterize, by means of a neural network (NN), objects to be processed visible in digital images of objects to be processed located in the object processing installation (FAC), the method comprising:
- a step (150) of generating digital training data (T.DAT) representative of simulated objects to be processed (S.IT) within the object processing installation (FAC), these data being intended for training the neural network, generated by means of a digital twin (NUM.TWIN) of the object processing installation, and associated with annotations (ANN) indicating at least one characteristic of the simulated objects to be processed (S.IT), the digital twin integrating a dynamic engine for simulating physical interactions between at least one object (IT) to be processed and at least one element (C1, C2, C3, M1, M2, M3, DZ, MZ) of the processing installation (FAC), so as to simulate movements of objects to be processed and interactions of these objects to be processed with the at least one simulated element of the processing installation;
- a step (160) of training the neural network comprising the provision of digital training data (T.DAT) as input to the neural network (NN) and during which the parameters of the neural network are adjusted on the basis of the digital training data;
- a step (170) of acquiring digital images of a real scene of the object processing installation (FAC) comprising at least one real object (IT) to be processed;
- a step (180) comprising the provision of data (R.DAT) generated from the digital images of the real scene as input to the neural network (NN), the neural network providing in return data representative of a characteristic (PGO) of the real object (IT) to be processed; and
- a step (190) comprising sending a control signal to an object processing device (MA) on the basis of the data representative of the characteristic of the at least one real object (IT) to be processed. Method (100) according to claim 1, in which the simulations of movements of objects to be processed and the simulations of interactions of the at least one object to be processed with the at least one element of the processing installation (FAC) comprise a simulation of a physical interaction between the at least one object to be processed and at least one conveyor (C1, C2, C3) of the installation (FAC), so as to simulate a configuration of the at least one object to be processed in a work zone (WZ) of the sorting installation. Method (100) according to claim 1 or 2, wherein the at least one simulated element is a ginning device (DE). Method (100) according to any one of claims 1 to 3, wherein the movements of simulated objects to be processed and the interactions of the at least one simulated object to be processed with the at least one simulated element comprise a simulation of an unloading of the objects to be processed in an unloading zone. Method (100) according to any one of claims 1 to 4, in which the at least one characteristic of the simulated objects to be processed is chosen from the position, the geometry and the orientation (P.G.O) of the simulated objects to be processed. Method (100) according to any one of claims 1 to 5, in which the digital training data (T.DAT) comprises at least one digital image of a simulated scene representing a working area (WZ) of the object processing installation (FAC). Method (100) according to claim 6, wherein the at least one digital image is generated so as to be as if it were obtained from digital images acquired by a digital camera (CAM) installed above the working area (WZ) of the actual processing installation. The method (100) of claim 6 or 7, wherein the digital training data further comprises a depth map of the simulated scene. The method (100) of claim 8, wherein the depth map is generated so as to be as if it were obtained from digital images acquired by a digital camera (CAM) installed above the working area (WZ) of the actual processing facility. Method (100) according to any one of the preceding claims 1 to 9, in which the digital twin integrates a dynamic engine for simulating physical interactions between simulated elements of the treatment installation, so as to reproduce movements and shocks between simulated objects to be treated and simulated elements of the treatment installation. Method (100) according to any one of the preceding claims 1 to 10, further comprising a step (140) of projecting textures from digital images (R.IMG) of real objects onto surfaces of the simulated objects to be processed (S.IT). Method (100) according to any one of the preceding claims 1 to 9, in which the digital twin of the object processing system is configured so as to generate digital images of homogeneous batches of objects to be processed comprising more than 40%, preferably more than 50%, preferably more than 80%, preferably more than 90% or 100% of objects belonging to the same type of objects to be processed. The method (100) of claim 12, wherein the type of objects is defined by a texture mapped onto surfaces of simulated objects. Treatment installation (FAC) comprising a calculator (CALC) configured to implement the method according to any one of the preceding claims 1 to 13 Installation according to claim 14, in which (i) the computer (Calc) is configured to control the manipulator arm (MA) and (ii) the computer (Calc) and the digital twin (NUM.TWIN) are interfaced in such a way that the at least one simulated element (M1, M2, M3) of the installation is controlled by the computer.