EP0706838B1 - Machine and method for sorting varied objects using at least one robotic arm - Google Patents

Abstract

The sorter consists of a conveyor (2) which carries the articles (E) to be sorted past a robot-controlled sorting post (3), where they are directed into appropriate containers (4). The sorting post comprises at least one robot arm (9) with a gripper (10), an artificial eye (8) which determines the position of articles on the belt and partially recognises their nature, and a complementary detection system (12) with one or more sensors to complete the identification of the articles. The data processing unit (15) processes the information from the artificial eye and sensors and causes the robot arm to pick up selected articles and drop them into the containers (4). The robot arm is of variable geometry, and the data processing unit has a computer which can recognise objects and their positions from the information provided and direct the arm to pick them up.

Description

The invention relates to a machine and a method. for sorting various objects, using at least one robotic arm, this machine can include advantageously a plurality of robotic arms arranged one after the other.

The object of the invention is to fully automate sorting of various heterogeneous objects, including conformation, weight and size may vary very sensitively.

According to a very interesting example of application, the invention can be advantageously implemented to perform automatic sorting, in particular for their recycling, discarded packaging and making part of household waste and similar waste, such as glass bottles or flasks, various plastic containers such as PVC (polyvinyl chloride), PET (polyester terephthalate), HDPE (high density polyethylene), food bricks made of mixed materials or cardboard, metal packaging, etc., the weight of which can vary from 10 grams to 1000 grams and which are brought in bulk, in varying proportions, to the sites sorting.

In this application, the waste stream of household packaging to be sorted, for example, comes from household waste for which a first sorting has been operated, either directly at the collection, or during previous sorting, according to known methods, in order to eliminate organic products and papers.

To sort a set of heterogeneous objects such that a flow of household packaging waste, we Often still applies the manual sorting solution. It is a very labor-intensive operation, which gives quality results often insufficient (which can be verified by the fact that recycling companies always do a additional refining sorting), and involves difficult working conditions for performers.

There are a number of sorting machines for household packaging waste, but these machines are, to date, mainly used after a first sorting operation, usually manual, so to remove the remaining impurities in a stream by elsewhere homogeneous.

These machines can be classified into two groups:
- A - The automated sorting machines, which consist in separating the objects, in order to make them pass in front of appropriate detectors, then in ejecting them selectively by means of pneumatic nozzles, in special tanks. A machine of this kind is, for example, described in documents WO-A-92/07332, and WO-A-92/01272.

In most such systems, the objects are first put in linear flow, then pass so individually in front of detectors which control or not ejection systems.

In some other cases, the objects are distributed on a conveyor, without being aligned, and subject to the same type of detectors as above but distributed across the plan. Ejection can also be done by nozzles.

• de ne pas permettre de trier des groupes d'objets constitués de matériaux différents en contact les uns avec les autres, de sorte qu'il est indispensable de disposer d'une machine d'égrenage permettant de singulariser les objets, en amont de la machine de tri;
     on conçoit qu'il en résulte une augmentation de l'encombrement de la machine de tri ;
• de ne pas être modulable, c'est-à-dire de ne pas permettre l'adjonction de modules de tri à une machine existante, par exemple, pour augmenter ou modifier sa capacité de traitement.

The disadvantages of these machines are:

• not to allow to sort groups of objects made up of different materials in contact with each other, so that it is essential to have a ginning machine making it possible to single out objects, upstream of the machine sorting;
  it is understood that this results in an increase in the size of the sorting machine;
• not to be modular, that is to say not to allow the addition of sorting modules to an existing machine, for example, to increase or modify its processing capacity.

In document WO-A-91/11885 is described a sorting machine with planar flow comprising transport means making it possible to convey products from a feeding station to a robotic sorting station, comprising at least one robot arm equipped with a gripping member and an artificial vision system for locating moving products and determining their orientation. However, this machine is specially designed to process products of the same or similar nature such as cakes, cookies, confectionery, etc. ; it is not arranged or configured to recognize heterogeneous objects and sort them according to their appearance, such as household waste, for example.
- B - Robotic sorting machines, intended to prevent manual operators from having direct contact with the objects to be sorted, for reasons of hygiene and safety.

These machines are generally suitable for a single category of waste, and are always ordered by an operator.

We know, for example, a sorting robot for hospital waste, which avoids the operator contamination. It is an arm of robot whose movements are calculated in a coordinate system Cartesian, operating in remote control mode by a operator. This sorting system does not include automatic recognition or localization of waste, and operates on whole bags.

We also know a robot intended to sort the food bricks and some bottles plastics, but this sorting system is also controlled by an operator, through a digital screen.

To date, none of the above machines have constitutes a very satisfactory overall solution to the sorting of discarded household packaging, mixed in any proportions.

The present invention aims to create a fully automatic sorting machine and process heterogeneous objects appearing in bulk on a conveyor belt and which are not affected shortcomings or disadvantages previously highlighted, known sorting machines and methods.

• le système de vision est capable de reconnaítre les caractéristiques visuelles des objets, et de les trier en fonction de ces caractéristiques ;
• le poste de tri robotisé comprend au moins un système de détection complémentaire comprenant au moins un capteur permettant de compléter l'identification des objets par la reconnaissance de leur matériau constitutif ;
• le poste de tri robotisé comprend des moyens de traitement des informations de localisation et d'identification fournies par le système de vision et par le système de détection complémentaire, et des moyens permettant le pilotage du bras robotisé assurant l'extraction sélective des objets localisés et identifiés, et leur dépose dans un bac approprié.

According to the invention, this object is achieved by means of a planar flow sorting machine, comprising transport means making it possible to convey said objects from a feeding station where the latter are discharged in bulk, up to at least a robotic sorting station, where they are distributed in appropriate containers, this robotic sorting station comprising at least one robot arm provided with a gripping member, and an artificial vision system making it possible to locate moving objects, this machine being remarkable in that:

• the vision system is capable of recognizing the visual characteristics of objects, and of sorting them according to these characteristics;
• the robotic sorting station comprises at least one complementary detection system comprising at least one sensor making it possible to complete the identification of the objects by the recognition of their constituent material;
• the robotic sorting station comprises means for processing the location and identification information supplied by the vision system and by the complementary detection system, and means allowing the control of the robotic arm ensuring the selective extraction of localized objects and identified, and place them in an appropriate container.

The automatic sorting process and machine according to the invention make it possible to make the recycling of household packaging waste which is currently very not very practical ; indeed, the only practical solution currently implemented for a set of waste heterogeneous, is manual sorting.

• l'automatisation totale du tri permet d'envisager de faire tourner les équipements en 2 × 8 heures, ou en 3 × 8 heures.
  Compte tenu du fait que la main d'oeuvre est actuellement le poste le plus coûteux dans l'opération de tri, cela permet de mieux amortir l'ensemble de l'installation de tri qui comprend notamment : le poste de tri, les postes d'amenée et d'évacuation des produits, les bâtiments ;
• la qualité de tri est meilleure et, surtout, a une meilleure répétabilité ; le tri d'affinage avant recyclage n'est alors plus nécessaire (alors qu'il est pratiqué actuellement pour pallier les erreurs humaines inévitables en tri manuel, surtout lors du tri de bouteilles fabriquées dans des matières plastiques différentes) ;
• le coût des postes de tri est réduit grâce à leur grande simplicité de construction ;
• les contraintes sanitaires et la pénibilité du travail impliquées par l'intervention d'opérateurs manuels dans le tri, sont supprimées.

The profitability provided by the invention compared to manual sorting is linked to the following factors:

• total sorting automation means that equipment can be run in 2 × 8 hours, or 3 × 8 hours.
  Given the fact that labor is currently the most expensive item in the sorting operation, this makes it possible to better amortize the entire sorting installation, which notably includes: the sorting station, the '' supply and evacuation of products, buildings;
• the sorting quality is better and, above all, has better repeatability; sorting before refining is no longer necessary (whereas it is currently practiced to overcome the inevitable human errors in manual sorting, especially when sorting bottles made from different plastics);
• the cost of sorting stations is reduced thanks to their great simplicity of construction;
• the health constraints and arduous work involved in the intervention of manual operators in sorting are eliminated.

• La possibilité de trier des groupes d'objets constitués de matériaux différents et se trouvant en contact les uns avec les autres ; selon l'invention, ces objets peuvent être pris et triés individuellement. Le système de vision artificielle permet d'isoler un objet dans un groupe d'objets en contact les uns avec les autres. Le bras automatisé ayant trié cet objet, le système de vision peut, par un processus itératif, trier de même les objets restant dans le groupe. Cette fonction, appelée "dégrappage", permet d'optimiser le tri.
• Un encombrement plus faible que celui des systèmes de tri conçus à ce jour. Ceci est dû à l'intégration, dans un seul poste de travail, de toutes les étapes importantes de tri : localisation des objets, préhension individuelle de ceux-ci, reconnaissance du matériau, et dépose dans le bac approprié, alors que, dans les systèmes connus, ces étapes sont réalisées en série par des dispositifs différents, chacun ajoutant son propre encombrement. Pour une capacité de traitement égale à celle de ces systèmes, la présente invention peut utiliser une pluralité de postes de tri robotisés, mais l'encombrement total de la machine revendiquée demeure inférieur à celui desdits systèmes.
• Une modularité découlant de sa conception sous forme de cellules de tri indépendantes qu'il est relativement facile d'ajouter les unes à la suite des autres.

In addition, compared to existing automatic sorting systems, the implementation of the machine and the method according to the invention, in particular provides the following important advantages:

• The possibility of sorting groups of objects made of different materials and being in contact with each other; according to the invention, these objects can be taken and sorted individually. The artificial vision system makes it possible to isolate an object from a group of objects in contact with each other. The automated arm having sorted this object, the vision system can, by an iterative process, similarly sort the objects remaining in the group. This function, called "unclipping", optimizes sorting.
• A smaller footprint than that of the sorting systems designed to date. This is due to the integration, in a single workstation, of all the important sorting steps: localization of objects, individual gripping of them, recognition of the material, and deposit in the appropriate bin, while in the known systems, these steps are carried out in series by different devices, each adding its own size. For a processing capacity equal to that of these systems, the present invention can use a plurality of robotic sorting stations, but the total size of the claimed machine remains less than that of said systems.
• A modularity stemming from its design in the form of independent sorting cells which it is relatively easy to add one after the other.

Modularity and reduced bulk allow progressive implementation thus reducing investment.

The above purposes, features and benefits and others will emerge better from the description which follows and attached drawings in which :

Figure 1 is a schematic view of a machine according to the invention comprising a plurality of stations robotic sorting.

Figure 2 is a perspective view and schematic, of a robotic sorting station individual.

Figure 3 illustrates, in a sectional view longitudinal, a very interesting embodiment of the complementary detection system, whereby said system is embedded inside the gripper of the robotic arm, the system being shown in its position before entering an object.

Figure 4 is a view similar to Figure 3 and showing the complementary detection system in its position resulting from the seizure of said object.

Reference is made to said drawings to describe a interesting, although by no means limiting, example of realization of the machine and implementation of the method according to the invention.

• un alimentateur 1 connu en soi, par exemple constitué par une trémie comportant un tapis de chargement 1a qui peut être formé par un convoyeur à bande sans fin ;
• une bande transporteuse 2 qui peut être avantageusement constituée par un transporteur à bande sans fin et sur l'extrémité de chargement de laquelle les objets à trier sont déversés, en vrac monocouche, par l'alimentateur 1 ;
• au moins un poste de tri robotisé 3 disposé sur le parcours du brin actif de la bande transporteuse 2 ; toutefois, de manière avantageuse, la machine comporte une pluralité de postes de tri (par exemple, cinq à sept postes de tri robotisé 3a, 3b, 3c, 3d, 3e, 3f,...) disposés les uns à la suite des autres le long de la bande transporteuse 2, de préférence à intervalles réguliers ;
• un ensemble de réceptacles de stockage 4, pour la réception des matériaux triés, disposés de chaque côté et le long de la bande transporteuse 2, à proximité du poste ou des postes de tri ; l'ensemble des réceptacles comprend également un réceptacle de rebut 4a, par exemple disposé à proximité de l'extrémité de sortie de la bande transporteuse; les réceptables 4 et 4a peuvent être constitués par des bacs de grande contenance, exécutés en matière plastique rigide ou autres matériaux présentant la rigidité et la robustesse souhaitables ;
• un module de supervision 5 commandant le poste de tri robotisé ou l'ensemble de la ligne des postes de tri robotisé.

This machine mainly comprises (figure 1):

• a feeder 1 known per se, for example constituted by a hopper comprising a loading belt 1a which can be formed by an endless belt conveyor;
• a conveyor belt 2 which may advantageously be constituted by an endless belt conveyor and on the loading end from which the objects to be sorted are poured, in single-layer bulk, by the feeder 1;
• at least one robotic sorting station 3 disposed on the path of the active strand of the conveyor belt 2; however, advantageously, the machine comprises a plurality of sorting stations (for example, five to seven robotic sorting stations 3a, 3b, 3c, 3d, 3e, 3f, ...) arranged one after the other along the conveyor belt 2, preferably at regular intervals;
• a set of storage receptacles 4, for receiving the sorted materials, arranged on each side and along the conveyor belt 2, near the station or sorting stations; the set of receptacles also comprises a waste receptacle 4a, for example disposed near the outlet end of the conveyor belt; receptables 4 and 4a can be made up of large capacity tanks, made of rigid plastic or other materials having the desired rigidity and robustness;
• a supervision module 5 controlling the robotic sorting station or the entire line of robotic sorting stations.

The conveyor belt 2 can be driven by means known per se, either at constant speed, either at controlled variable speed or with stops intermittent. It can be formed continuously, or consisting of several successive sections independent. In all cases, the sorting station robotized or each robotic sorting station must at least know the law of motion of the band. This law of movement is controlled by the supervision module or supervisor 5 which includes a computer or PLC programmable that receives information from belt or conveyor belt 2 (measurement of speed) depending on the occupancy rate of the robots the line and acts on the drive motor (s) 6 of said band.

In other words, the conveyor belt 2 can be advantageously subject to the operation of the robotic sorting, so as to optimize the sorting rate, the speed of movement of said conveyor belt being accelerated when the station vision system robotic sorting no longer or very little objects, or, conversely, slowed down when objects transported by the band are numerous.

The advantage of this management is to adapt waste stream automatically to the capacity of processing of all robots on the same line. For example, for simplicity and reliability, this speed can be changed at intervals of discrete times, instead of making it evolve continuously. The band movement law is then made up of successive speed steps.

Intermittent stop mode of travel consists in making the robots work on the belt to stopping, which simplifies and facilitates docking of localized objects. The band is only activated by moments, so as to empty the remaining objects from the end of the chain to the receptacle of scrap 4a, and to simultaneously recharge a batch of objects new to sort.

• un portique 7 enjambant la bande transporteuse 2 et supportant les différents équipements disposés au dessus de cette dernière ;
• un système de vision artificielle permettant la localisation des objets et la reconnaissance partielle de leur matériau constitutif, ce système de vision, désigné dans son ensemble par la référence 8, comprenant son éclairage spécifique, une caméra, de préférence une caméra couleur, et un système électronique d'acquisition ; ce système de vision artificielle est capable de localiser les objets, de reconnaitre leurs caractéristiques visuelles, de les trier en fonction de celles-ci et, surtout, de réaliser toutes ces opérations sur des objets mal individualisés, c'est-à-dire se trouvant en contact les uns avec les autres ;
• un bras robotisé 9 à géométrie variable et à au moins trois degrés de liberté, par exemple à structure sphérique, dont la particularité est que le foyer de la caméra du système de vision 8 se trouve disposé fixement au centre du repère sphérique du robot ; un tel bras robotisé articulé ou manipulateur à coordonnées sphériques (deux rotations concourantes et une translation vers l'objet à saisir) possédant un système de vision fixe en son centre, étant décrit dans le document EP-A-0270469 ; le bras robotisé 9 et le système de vision 8 sont portés par la poutre horizontale 7a du portique 7 disposée au-dessus de la bande transporteuse 2. De manière préférée, le bras robotisé est centré par rapport à l'axe de défilement de la bande transporteuse 2 ;
• un préhenseur 10 constitué par l'extrémité libre réalisée sous forme de trompe, du bras robotisé et permettant de saisir les objets localisés par le système de vision 8, c'est-à-dire l'un quelconque des objets déposés sur la bande transporteuse 2, dans toutes les positions accessibles au bras robotisé 9 ; ce préhenseur est, par exemple, formé par un dispositif pneumatique fonctionnant en aspiration (ventouse) et/ou en soufflage ; ce dispositif, connu en soi, permet de saisir un objet préalablement identifié, par aspiration, et ensuite de l'éjecter, par soufflage, dans le réceptacle approprié 4, en fonction de sa nature ;
• au moins un capteur 11, monté sur le bras robotisé 9, permettant de contrôler la saisie correcte des objets par le préhenseur 10 ; il peut s'agir, par exemple, d'un capteur de dépression monté sur le circuit d'aspiration ;
• au moins un système de détection complémentaire 12 comportant au moins un capteur permettant la confirmation de la nature des matériaux des objets saisis par le préhenseur 10, par exemple un capteur permettant d'analyser les spectres électromagnétiques spécifiques émis par les différents matériaux constitutifs desdits objets (dans le domaine proche de l'infrarouge, ultraviolet, ou rayons X, etc.), ou un capteur apte à détecter la présence de métaux, ou un capteur apte à mesurer le poids des objets : ce système de détection complémentaire peut être ou ne pas être embarqué sur le robot ;
• une armoire de commande 13 permettant de contrôler le système de vision 8, le bras robotisé 9, le préhenseur 10 et les capteurs 11 et 12 ;
• une armoire de puissance 14 renfermant les circuits de commande et les circuits de puissance, cette armoire fournissant l'énergie électrique au système de vision 8, au bras robotisé 9, au préhenseur 10 et aux capteurs 11 et 12 ;
• un module logiciel 15 permettant le traitement des informations transmises par le système de vision 8 et par le système de détection complémentaire 12 ;
• un module logiciel 16 permettant le pilotage du bras robotisé 9 et du préhenseur 10, suivant les informations fournies par le module logiciel 15 ;
• des goulottes d'éjection 17, par exemple supportées par le châssis 18 soutenant la bande transporteuse 2, chacune d'elles aboutissant dans l'un des réceptacles de stockage 4 des objets triés.

Each robotic sorting station 3 comprises the following constituent elements (FIG. 2):

• a gantry 7 spanning the conveyor belt 2 and supporting the various equipment disposed above the latter;
• an artificial vision system allowing the localization of objects and the partial recognition of their constituent material, this vision system, generally designated by the reference 8, comprising its specific lighting, a camera, preferably a color camera, and a system acquisition electronics; this artificial vision system is capable of locating objects, of recognizing their visual characteristics, of sorting them according to these and, above all, of carrying out all these operations on poorly individualized objects, that is to say finding in contact with each other;
• a robotic arm 9 with variable geometry and at least three degrees of freedom, for example with a spherical structure, the peculiarity of which is that the focus of the camera of the vision system 8 is fixedly disposed in the center of the spherical coordinate system of the robot; such an articulated robotic arm or manipulator with spherical coordinates (two concurrent rotations and a translation towards the object to be grasped) having a fixed vision system at its center, being described in document EP-A-0270469; the robotic arm 9 and the vision system 8 are carried by the horizontal beam 7a of the gantry 7 disposed above the conveyor belt 2. Preferably, the robotic arm is centered relative to the axis of travel of the belt conveyor 2;
• a gripper 10 constituted by the free end produced in the form of a trunk, of the robotic arm and making it possible to grasp the objects located by the vision system 8, that is to say any one of the objects deposited on the conveyor belt 2, in all the positions accessible to the robotic arm 9; this gripper is, for example, formed by a pneumatic device operating in suction (suction cup) and / or in blowing; this device, known per se, makes it possible to grasp an object previously identified, by aspiration, and then to eject it, by blowing, into the appropriate receptacle 4, depending on its nature;
• at least one sensor 11, mounted on the robotic arm 9, making it possible to control the correct gripping of the objects by the gripper 10; it may be, for example, a vacuum sensor mounted on the suction circuit;
• at least one complementary detection system 12 comprising at least one sensor allowing confirmation of the nature of the materials of the objects seized by the gripper 10, for example a sensor making it possible to analyze the specific electromagnetic spectra emitted by the various materials constituting said objects ( in the near infrared, ultraviolet, or X-ray domain, etc.), or a sensor capable of detecting the presence of metals, or a sensor capable of measuring the weight of objects: this complementary detection system may or may not not be on board the robot;
• a control cabinet 13 for controlling the vision system 8, the robotic arm 9, the gripper 10 and the sensors 11 and 12;
• a power cabinet 14 containing the control circuits and the power circuits, this cabinet supplying electrical energy to the vision system 8, to the robotic arm 9, to the gripper 10 and to the sensors 11 and 12;
• a software module 15 allowing the processing of the information transmitted by the vision system 8 and by the complementary detection system 12;
• a software module 16 enabling the robotic arm 9 and the gripper 10 to be controlled, according to the information provided by the software module 15;
• ejection chutes 17, for example supported by the frame 18 supporting the conveyor belt 2, each of them ending in one of the storage receptacles 4 of the sorted objects.

According to another characteristic arrangement of the invention, the machine is provided with a plurality of complementary sensors 12, allowing identification automatic objects by confronting information provided by at least one of them with those provided by the identification system visual 8, triggered as needed, and taken into account according to predefined priorities.

• a) d'accroítre la qualité de détection (par exemple en évitant les étiquettes) ;
• b) de mobiliser les capteurs complémentaires pendant une fraction de temps moins importante, ce qui permet un multiplexage sur plusieurs cellules.

In this case, according to the method of the invention, it is the artificial vision device which pre-selects, as necessary, the portion of the object where the measurement of the complementary sensor or sensors is to be carried out. This feature allows:

• a) to increase the quality of detection (for example by avoiding labels);
• b) to mobilize the complementary sensors for a smaller fraction of time, which allows multiplexing on several cells.

According to an interesting characteristic arrangement of the invention, the complementary detection system 12 is embedded inside the gripper 10 of the arm robotized 9, so as to follow all the movements of said arm and allow to define the nature of objects, as soon as they are entered.

• un capteur inductif 12A réagissant en présence de métal et permettant la détection des objets métalliques (ferreux et non ferreux) ;
• un capteur de chocs 12B permettant la détection des objets massifs (verre, faïence, objets métalliques de forte épaisseur, etc.) ; ce capteur est constitué par un accéléromètre, connu en soi, émettant des signaux proportionnels à l'accélération de l'objet sur lequel il est posé, cet accéléromètre permettant, dans l'application à la présente invention, l'évaluation de la dureté des objets saisis, comme expliqué dans la suite du présent exposé.

According to the advantageous embodiment shown in FIGS. 3 and 4, the complementary detection system 12 includes two specific sensors, namely:

• an inductive sensor 12A reacting in the presence of metal and allowing the detection of metallic objects (ferrous and non-ferrous);
• a shock sensor 12B allowing the detection of massive objects (glass, earthenware, very thick metallic objects, etc.); this sensor is constituted by an accelerometer, known per se, emitting signals proportional to the acceleration of the object on which it is placed, this accelerometer allowing, in the application to the present invention, the evaluation of the hardness of the seized objects, as explained later in this presentation.

Inductive sensor 12A and shock sensor 12B are placed at opposite ends of a support elongated shape, for example consisting of a rod tubular 19 passing through a guide 20 fixedly installed at the interior of the rigid tube constituting the gripper 10. The inductive sensor 12A is arranged at the end front of rod 19, while shock sensor 12B is arranged at the rear end of the latter. The sensors 12A and 12B constitute, with the rod 19, a assembly or mobile equipment mounted with an ability to axial movement in the gripper 10.

A means tends to push, permanently, the whole mobile 12A-12B-19 towards the active end of the gripper 10. This means is, for example, constituted by a spring 21 acting in compression and wedged, by through its opposite ends, on the one hand, against a bulge presented by the anterior part of the rod 19, and, on the other hand, against the end front of guide 20. A mechanical stop 22 allows stop the mobile assembly in its extreme position progress (Figure 3).

The mechanical stop 22 is fixed on the part posterior of the rod 19 of the mobile assembly. She is shaped to protect the shock sensor 12B and it extends in front of the latter. In position advancement of the mobile assembly, it stumbles against the posterior end of the guide 20.

The cables 23 from the inductive sensor 12A and the shock sensor 12B pass through the rod 19, to exit, together, through a sealing piece 24 equipping the rear end of said rod.

The signals emitted by the sensors 12A and 12B are transmitted to shaping means known per se, allowing them to be used by a system identification.

The 12B accelerometer mounted on the mobile assembly 12A-12B-19 transmits the value of the acceleration undergone by said mobile assembly to an electronic card of conversion. This card transforms the input signal analog from the accelerometer as an output signal binary or analog usable by the system identification. This output signal has the value 0 (or 1) when the input signal is weak and for value 1 (or 0 respectively), when the signal entry is strong. The electronic card has a system for setting the minimum value of the input signal giving an output signal equal to 1.

The sorting process and the operation of the machine according to the invention, in considering a sorting installation comprising a plurality of robotic sorting stations.

The products to be sorted are poured in bulk into the hopper of feeder 1 and brought, by the conveyor belt unloading the latter, on the extreme part upstream of the moving conveyor belt 2. The supervision module 5 starts the cycle simultaneously for all robotic sorting stations on the machine.

Image acquisitions are taken at intervals until the vision system 8, controlled by software module 15, signals the presence an object, locates it, and determines if that object is to seize. If the object is to be grasped, the position of its center of gravity (X, Y) in the field of vision, as well that the position of the conveyor belt 2 when image acquisition are provided by the module software 15 to software module 16. The software module 16 then gives the robotic arm 9 the order to carry out the sorting cycle. If the object should not be seized, it continues its journey on the conveyor belt 2, to the next robotic sorting station or, depending on the case, up to the waste receptacle 4a.

The sorting cycle breaks down as follows: next: the robotic arm 9 deploys, its end gripping orients itself towards the center of gravity of the object identified and its gripper 10 is activated (by example by switching on the suction). he advances towards the object, in order to seize it. If the object is moving, the arm is controlled by the module software 16, so as to anticipate the position said object during docking, depending on the speed of movement of the conveyor belt. he stops, either at the end of its travel, or when the control sensors 11 signal the presence of object or obstacle. The sensors 11 make it possible to determine if the arm took an object or encountered a barrier. If it is indeed an object, he grabs it through its gripper 10, then active the analysis sensor (s) 12 to determine the material of which it is made.

When entering, the anterior end of the mobile assembly 12A-12B-19 housed in the gripper 10 comes into contact with object E (Figure 3).

Under the action of the depression prevailing in the gripper 10, the object E is sucked up and driven towards the rear, compressing the flexible end 10A of said gripper and causing the set back mobile 12A-12B-19.

The inductive sensor is continuously analyzed by the identification system.

Upon contact with a metallic object, the inductive sensor 12A signals the presence of metal.

• dans le cas d'objets massifs non déformables ou peu déformables, l'énergie de l'objet est intégralement transmise à l'ensemble mobile qui subit une accélération instantanée importante ;
• par contre, pour les objets légers déformables (en plastique par exemple), la déformation de l'objet, sous l'action du ressort 21, absorbe une partie de l'énergie dudit objet ; l'ensemble mobile 12A-12B-19 subit donc une accélération instantanée faible.

At the time of contact between the object E and the mobile assembly 12A-12B-19, the latter undergoes an acceleration due to the relative movement of said object relative to the gripper. The maximum value of instantaneous acceleration depends largely on the nature of the object. Indeed :

• in the case of solid non-deformable or slightly deformable objects, the energy of the object is fully transmitted to the mobile assembly which undergoes a significant instantaneous acceleration;
• on the other hand, for light deformable objects (in plastic for example), the deformation of the object, under the action of the spring 21, absorbs part of the energy of said object; the mobile assembly 12A-12B-19 therefore undergoes a slight instant acceleration.

As previously mentioned, the shock sensor 12B transmits the value of the acceleration undergone by the mobile assembly, to an electronic card conversion.

Depending on the material identified, the arm turns and moves over the chute 17 leading to the particular receptacle 4 intended for the reception of objects made from this material. The deposit of the object removed from the planar flow, in the appropriate chute 17, can be done by simple release of said object. This removal can also be advantageously produced by blowing with the gripper 10. The advantage of removal by blowing not to require an approach movement of the arm in direction of the drop zone in chute 17, ni, therefore, backward movement, which allows thus saving time in the cycle. As shown in Figure 2, the arrangement of the troughs 17 can be optimized to avoid any extension of the arm robotized before removal.

When the robotic arm has not taken an object from the following its deployment, it returns to the position of fixed rest folded. It waits, from the software module of piloting 16, the order to continue on the next object, either from this rest position or from from its removal position.

During the seizure of an object, and its deposit in the suitable sorting receptacle 4, by the robotic arm 9, a new image acquisition is carried out by the vision system 8. This image is processed by the processing software module 15, in masked time, with the cycle of the robotic arm 9, that is to say while the seized object is transported to the chute 17 appropriate. As soon as the object is removed transport to the chute, the software module of piloting 16 is thus able to give a new robotic arm sort order 9.

When the conveyor belt 2 moves from intermittently, localization, E objects are identified and entered during periods of stoppage of the said strip, which facilitates and simplifies docking of objects, as previously pointed out.

As previously stated, the machine or sorting facility can have multiple arms robotic 9 placed in series one after the other others, along the conveyor belt, which provides several additional benefits in the installation configuration.

It is possible to adapt the number of arms robotic 9 at the desired rate, fairly progressive. If, for example, an installation uses five robotic arms for 100% capacity, one 120% capacity requires only the addition of a sixth robotic arm. Robotic sorting stations can thus be produced in the form of modules can be added one after the other.

In the event of a temporary arm failure robotic, the installation continues to operate, but at reduced capacity.

We can possibly specialize more or less robotic arms, depending on the type of waste, or categories of objects, each sorting station being then assigned to the recognition and removal of objects belonging to a specific category of objects (for example, PVC bottles) and being programmed in result. For example, five robotic arms 9 each associated with 4 discharge channels 17, allow up to twenty output categories, number very high given the size of the installation. This specialization can be dynamic, i.e. depending on the composition average observed or known in advance of a batch of donated waste.

We can share, by several sorting, sensor acquisition / processing units 12, to significantly reduce costs, the analysis performed by these sensors taking only a small fraction of the work cycle (typically less than 50 ms), and the preselection made by the group robot-camera avoiding making these sensors work for the entire duration of the cycle.

In other words, when one of the sensors of the complementary detection system 12, is constituted by a spectroscopic analysis sensor, it can be installed so that it can be multiplexing technique, measurements on several robotic cells.

Claims (27)

1. Machine with planar flow for automatically sorting various objects such as, for example, household packaging waste, having transfer means (2) enabling said objects (E) to be routed from a feeder station where they are poured out in bulk, as far as at least one robotised sorting station (3) where they are distributed into suitable receptacles (4), said robotised sorting station (3) comprising at least one robot arm (9) equipped with a gripping means (10) and a machine vision system (8) which makes it possible to locate the objects deposited on the transfer means (2), characterised in that:

   the machine vision system (8) is able to recognise the visual features of the objects, and to sort them according to these features;

   the robotised sorting station comprises at least one complementary detection system (12) comprising at least one sensor (12A, 12B) which makes it possible to complete the identification of the objects by recognising the material of which they are constituted;

   the robotised sorting station comprises means (15) for processing the locating and identifying information supplied by the machine vision system (8) and by the complementary detection system (12) and means (16) enabling the robotised arm (9) to be driven so as to selectively withdraw the objects located and identified, and enabling them to be placed in a suitable receptacle (4).
2. Automatic sorting machine according to claim 1, characterised in that the robot arm (9) is constituted by an articulated robotised arm with variable geometry and spherical coordinates (two concurrent rotations and translation towards the object to be grasped) possessing a machine vision system (8) in its centre.
3. Automatic sorting machine according to either of claims 1 and 2, characterised in that the means for processing locating and identifying information (15) comprise a computer equipped with software able to locate the objects, to recognise their visual features, and above all to carry out all these operations on poorly separated objects, that is to say objects that are in contact with one another.
4. Automatic sorting machine according to any of claims 1 to 3, characterised in that it has a computer enabling the robotised arm (9 - 10) to be driven via software (16) which processes the information provided by the locating and identifying software (15).
5. Automatic sorting machine according to any of claims 1 to 4, characterised in that the complementary detection system (12) comprises a plurality of sensors (12A, 12B) making it possible to automatically identify objects by comparing the information provided by at least one of them with the information provided by the machine vision system (8).
6. Automatic sorting machine according to any of claims 1 to 5, characterised in that it comprises at least one sensor (11) making it possible to check that the gripper (10) of the robotised arm (9) grasps the objects correctly.
7. Automatic sorting machine according to any of claims 1 to 6, characterised in that the gripper (10) of the robotised arm (9) is constituted by a pneumatic device enabling an object (E) to be grasped by suction and to be ejected by being blown into a receptacle (4) appropriate to the nature of said object.
8. Automatic sorting machine according to any of claims 1 to 7, characterised in that the complementary detection system (12) is housed inside the gripper (10) of the robotised arm, thereby enabling the nature of the grasped objects (E) to be known as soon as they come into contact with the operational end (10a) of the gripper (10).
9. Automatic sorting machine according to claim 8, characterised in that the complementary detection system (12) loaded inside the gripper (10) comprises: an inductive sensor (12A) for recognising metal objects; a shock sensor or accelerometer (12B) for assessing the hardness of the object; said system further comprising means for shaping the signals output by the sensors (12A, 12B) in order to render them usable by an identification system.
10. Automatic sorting machine according to claim 9, characterised in that the inductive sensor (12A) and the shock sensor (12B) are respectively arranged at the front end and in the posterior portion of an axially movable support of elongated shape (19), one means, for example constituted by a spring (21) operating in compression, seeking to constantly push the movable assembly (12A, 12B, 19) towards the operational end (10a) of the gripper (10).
11. Automatic sorting machine according to any of claims 1 to 10, characterised in that it comprises means which enable the conveying belt (2) to be displaced either at a constant speed, at a controlled variable speed, or intermittently.
12. Automatic sorting machine according to any of claims 1 to 11, characterised in that the conveying belt (2) upon which the objects that are to be sorted (E) are travelling is matched to the operation of the robotised arm, via a servo system known per se.
13. Automatic sorting machine according to any of claims 1 to 12, characterised in that it comprises a plurality of robotised sorting stations (3a, 3b, 3c, 3d, 3e, 3f, ...) fitted one after the other on the same sorting belt (2), thereby enabling them to adapt to variable flow rates or to allow each sorting station to be specialised for the identification and removal of the objects belonging to a specific category.
14. Automatic sorting machine according to either of claims 1 to 13, characterised in that each sorting station comprises a gantry support (7) straddling the conveying belt (2) constituting the means for transporting the objects (E) to be sorted, and in that the machine vision system (8) and the robotised arm (9) are carried by the horizontal beam (7a) of this gantry support (7).
15. Automatic sorting machine according to either of claims 13 and 14, characterised in that it has one or more complementary sensors (12) installed in a way that allows them, by a multiplexing technique, to perform measurements on the entirety of the robotised cells (3a, 3b, 3c, 3d, 3e, 3f) of said machine.
16. Method for automatically sorting various objects such as, for example, household packaging waste, characterised by the following steps:

    the objects (E) are poured out in bulk onto a conveying belt (2) and said objects are routed as far as at least one robotised sorting station (3) comprising a machine vision system (8), at least one complementary detection system (12) comprising at least one sensor (12A, 12B) and a robotised arm (9);

    the machine vision system (8) partially locates and identifies the objects;

    the objects are identified in a complementary manner by one or more sensors (12A, 12B) of the complementary detection system (12);

    the locating and identifying information provided by the machine vision system (8) and by the sensor or sensors (12A, 12B) of the complementary detection system (12) is analysed by a first information processing system (15);

    using data transmitted by the first information system (15), a second information processing system (16) drives the robotised arm (9), which grasps a located and identified object and deposits it in a suitable receptacle (4).
17. Method for automatic sorting according to claim 16, characterised in that the complementary detection system (12) is accommodated in the gripper (10) of the robotised arm (9) in the form of a movable unit (12A, 12B, 19) arranged so as to come into contact with the objects (E) when they are grasped, such that the sensor or sensors (12A, 12B) of said detection system is or are operational and that the nature of an object (E) is defined as soon as said movable unit comes into contact with said object.
18. Method for automatic sorting according to either of claims 16 and 17, characterised in that the identification of a located object partly takes place while the object, which has previously been grasped and identified, is being deposited in the appropriate sorting receptacle (4).
19. Method for automatic sorting according to any of claims 16 to 18, characterised in that the automatic identification of the objects is done by comparing the information provided by a plurality of sensors which are brought into action as required, and taken into account in accordance with predefined priorities.
20. Method for automatic sorting according to claim 19, characterised by carrying out, via the machine vision system (8) and according to necessity, a preselection of that portion of the objects (E) upon which the measurement by the complementary sensor or sensors (12) must be carried out.
21. Method for automatic sorting according to any of claims 16 to 20, characterised in that while sorting is underway the conveying belt (2) travels either at a constant speed or at a controlled variable speed.
22. Method for automatic sorting according to any of claims 16 to 20, characterised in that the conveying belt (2) receiving the objects to be sorted is set in motion intermittently, and that the objects are located, identified and grasped during the periods when said belt is at a standstill.
23. Method for automatic sorting according to claim 21, characterised in that the travelling speed of the conveying belt (2) is matched to the operation of the robotised arm (9 - 10).
24. Method for automatic sorting according to any of claims 16 to 23, characterised in that the located and identified objects (E) are grasped by the robotised arm (9 - 10) using suction, and are then ejected by blowing them into the appropriate receptacles (4).
25. Method for automatic sorting according to any of claims 16 to 24, characterised by arranging a plurality of robotised sorting stations (3a, 3b, 3c, 3d, 3e, 3f) one after the other.
26. Method for automatic sorting according to claim 25, characterised in that each robotised sorting station (3a, 3b, 3c, 3d, 3e, 3f) is programmed to carry out the recognition and sorting of the objects belonging to a specific category of objects.
27. Method for automatic sorting according to either of claims 25 and 26, characterised in that the measurements taken on a plurality of robotised cells (3a, 3b, 3c, 3d, 3e, 3f) are transmitted, by multiplexing, to one or more complementary sensors (12).

Images