The process of interaction with the subject

Abstract

A process of interaction with an object (1a) is provided comprising a robot (2) with an end effector (21), a driver (22) for the end effector (21) and a sensor (23) for acquiring at least one environment or interaction parameter; a computer (5) for controlling the robot (2); and an instruction block (3) configured to command the robot (2) to move the end effector (21) according to instructions defined by the instruction block.

Description

[Technical Field]

[0001] The present invention relates to a process of interaction with an object of the kind specified in the preamble of claim 1.

[0002] In particular, the present invention relates to a process configured to use robots for handling objects, preferably in an automated production assembly line. [Background technology]

[0003] As is well known, automated production assembly lines are particularly flexible thanks to the use of robots that are able to perform different tasks and handle products of different shapes and densities.

[0004] Such robots are now used to move objects between different stations, for example between warehouses and workstations, and / or to perform tasks such as assembly and welding.

[0005] They are programmed to faithfully perform repetitive tasks with high precision, their actions determined by software that specifies the direction, acceleration, speed, and distance of a series of coordinated movements.

[0006] These movements are defined by an operator who, for each movement, must define a command, therefore an instruction to be given to the robot to enable it to carry out the desired movement.

[0007] The known techniques described have some important drawbacks.

[0008] In particular, known handling procedures require the use of specific programming languages ​​to command the robot for each single action of each new task, resulting in complex management of the manufacturing process and frequent errors due to incorrect identification of the correct plan to be executed.

[0009] To solve this problem, a new process for defining instructions has been devised, and new commercial robots have been equipped with graphical user interfaces (GUIs) that allow operators to program them in a short amount of time.

[0010] Although this solution has significantly accelerated programming, it has not resulted in a significant improvement. As a result, known object handling procedures are currently particularly complex, not easy to use, relatively high cost, and / or involve the use of complex and expensive robotics.

[0011] Another drawback is the lack of flexibility of known handling processes, which makes it necessary to almost constantly reprogram them, since the robot is unable to adapt to any changes in the movements to be performed and / or the objects to be handled. Summary of the Invention

[0012] In this context, the technical problem underlying the present invention is to devise a procedure for interaction with an object that makes it possible to substantially eliminate at least some of the aforementioned drawbacks.

[0013] In the context of the above mentioned technical problems, it is an important object of the present invention to obtain an easily programmable and therefore highly flexible interaction procedure with an object.

[0014] The technical problem and the stated object are achieved by a process of interaction with an object as claimed in the attached claim 1. Examples of preferred embodiments are set out in the dependent claims. [Brief explanation of the drawings]

[0015] The features and advantages of the present invention will become apparent from the following detailed description of preferred embodiments of the invention, taken in conjunction with the accompanying drawings.

[0016] [Figure 1]1 shows, in a reduced scale, a device configured to implement a process for interacting with an object according to the invention; [Figure 2a] 2 shows the assembly of the device of FIG. 1 at a reduced scale. [Figure 2b] 2a and 2b show, in reduced scale, the assembly of FIG. 2a in different uses; [Figure 3] 1 shows, in a reduced scale, an environment for carrying out a process of interacting with an object according to the invention. [Figure 4] 1 shows a schematic of a process for interacting with an object according to the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0017] In this document, measurements, values, shapes, and geometric references (e.g., perpendicularity and parallelism), when associated with words like "about" or other similar terms such as "approximately" or "substantially," should be considered to exclude measurement errors or inaccuracies due to production and / or manufacturing tolerances, and in particular to exclude slight deviations from the value, measurement, shape, or geometric reference with which it is associated. For example, when associated with a value, these terms preferably indicate deviations of 10% or less of the value.

[0018] Furthermore, when used, terms such as "first," "second," "higher," "lower," "primary," and "secondary" do not necessarily specify a priority of order, relationship, or relative position, but may be used merely to clearly distinguish between different components thereof.

[0019] Measurements and data reported in this text should be considered as having been carried out to the International Standard Atmosphere ICAO (ISO 2533:1975) unless otherwise stated.

[0020] Unless otherwise indicated, as a consequence of the following discussion, terms such as "processing," "calculating," "determining," "computing," or similar terms refer to operations and / or processes of a computer or similar electronic computing device that manipulates and / or transforms data represented as physical quantities, such as electronic quantities in the registers and / or memory of a computer system, into other data similarly represented as physical quantities in the computer system, registers, or other storage, transmission, or information display device. With reference to the drawings, processes that interact with the subject matter of the present invention are generally designated by the numeral 1.

[0021] The process 1 is configured to recognize and interact with objects 1a within an environment 10, such as a house or a flat, and preferably within a work / industrial environment 10, such as a warehouse or a production line, and thus to recognize, interact with and thus move / handle objects 1a.

[0022] The surrounding environment 10 may define a walkable area 10a.

[0023] The ambient environment 10 may comprise at least one room 11. In particular, it may comprise a plurality of rooms 11 and at least one passage 12 between each of the rooms.

[0024] The interaction process 1 may comprise at least one robot 2 .

[0025] The robot 2 may be configured to perform an interaction with at least one object 1 a. The robot 2 may include at least one end effector 21 for grasping the at least one object 1 a.

[0026] The end effector 2 is configured to interact with objects 1 a among a plurality of objects 1 a that preferably differ from one another in weight and / or shape, for example, it is configured to interact with only one object 1 a at a time and precisely manipulate / perform a grasp thereof.

[0027] The end effector 2 may comprise at least two mechanical fingers movable relative to one another to perform the grasp, preferably it comprises more than two fingers, in particular those identifiable in robotic hands such as those described in US2019269528 and US2018311827.

[0028] The robot 2 may include, for each end effector 21, a driver 22 for the end effector 21.

[0029] The driver 22 may be configured to move the end effector 21 relative to the object 1 a and / or the surrounding environment 10 .

[0030] The driver 22 may comprise a robotic arm, which may comprise one or more rigid bodies 221, optionally identifiable with nested profiles, and one or more joints 222 suitable for moving the rigid bodies 221, preferably independently.

[0031] The joints 222 can be configured to rotate the rigid bodies 221 relative to one another by changing the angle subtended between two adjacent bodies 221 .

[0032] Preferably, the joints 222 move the rigid body 221 according to inverse kinematics or direct kinematics. Thus, even if not explicitly stated, it is provided that each movement of the robot 2, and therefore the commands defining said movement, are determined according to inverse or direct kinematics.

[0033] The term inverse kinematics defines the trajectory in motion space of the path of the end effector 21. Thus, the velocity and acceleration of each joint 222 is determined to comprise the path 21 of the end effector.

[0034] The term direct kinematics identifies the calculation of a trajectory in space, in which the position, velocity, and acceleration of each joint 222 are determined, rather than the path of the end effector 21. Optionally, the path of the end effector 21 is the result of the position, velocity, and acceleration of each joint 222.

[0035] Each joint 222 may be motorized, and in particular may comprise a servo motor.

[0036] The driver 22 may comprise a displacement means 223 configured to move the robot 2 along the walkable plane 10a.

[0037] The displacement means 223 may be motorized, for example they may comprise at least one crawler or wheel.

[0038] The robot 2 may be equipped with sensors 23 for acquiring at least one suitable environmental parameter and / or for the interaction of the robot 2 with the object 1a.

[0039] The sensor 23 may comprise one or more sensors each configured to acquire parameters selected from ambient environment parameters and interaction parameters, and preferably configured to acquire at least one ambient environment parameter and at least one interaction parameter.

[0040] The expression "ambient environment parameters" identifies parameters that are external to the robot 2 and therefore not relevant to its operation. Ambient environment parameters may be physical conditions specific to the environment 10 (e.g., temperature, humidity, brightness) and / or characteristics of the object 1a (e.g., shape and / or weight).

[0041] The expression "interaction parameters" identifies parameters related to the behavior of the robot 2 when interacting / manipulating with the object 1 a, which may be identified for example in the gripping position and / or force of the end effector 21 or in the contact temperature of the end effector 21 with the object 1 a.

[0042] In the case of acquiring at least one ambient environment parameter, the sensing means 23 may comprise one or more sensors selected from a thermometer, a photodetector, a hygrometer, or a means for imaging the object 1a, such as a camera.

[0043] In the case of an interaction parameter, the sensor 23 may comprise one or more sensors selected from a thermometer, a piezoelectric, an encoder for each joint 222 .

[0044] The sensors 23 can be configured to acquire the movements of the robot 2 so as to make it possible to determine the commands necessary to determine the movements of the robot.

[0045] The robot 2 may comprise a card 24 for controlling the robot 2 , in particular at least the end effector 21 and the driver 22 .

[0046] The card 24 may be in data communication with the sensor 23 .

[0047] The interaction procedure 1 may comprise an instruction block 3 configured to define and send instructions to the robot 2 for moving an end effector 21 (preferably a driver 22) and thus interacting with one or more objects 1a.

[0048] The command block 3 can be in data communication with the robot 2, in particular with the computer 24.

[0049] Block 3 may be configured to locate object 1 a in the surrounding environment 10 and to define and send to robot 2 the commands necessary to perform an interaction (handle, grasp, and / or move) of object 1 a. For example, block 3 may be configured to instruct robot 2 to locate object 1 a if present in the surrounding environment 10, grasp object 1 a, perform one or more actions with respect to object 1 a, and then store it at the location where it is located in the surrounding environment 10.

[0050] The command block 3 may comprise a computer or other device configured to allow an operator to input such commands.

[0051] Alternatively, the instruction block 3 may be at least partially wearable (FIG. 1) to enable the operator to simulate actions and then send said commands to the robot 2 according to the operator's movements. It may comprise a visualizer 31 configured to present to the operator at least some of the parameters acquired by the sensors 23; and acquisition means 32 configured to acquire the movements performed by the operator and to send to the robot 2 the commands necessary to enable the robot to repeat said movements.

[0052] The acquisition means 32 may comprise a sensorized glove configured to detect the operator's hand movements and send commands to the robot 2, for example, to enable the hand movements to be executed by the end effector 21.

[0053] Alternatively, the means 32 may comprise a sensorized suit configured to detect the movements of the operator and then send commands to the robot 2 so that said movements can be carried out.

[0054] The instruction block 3 may include at least one camera 33, and the viewer 31 may include a screen (eg, virtual glasses) for viewing the camera image.

[0055] The camera 33 may be integrated into the robot 2 (FIG. 1).

[0056] Alternatively, the cameras 33 may be integrated into the rooms 10. Preferably, the command block 3 comprises, for each room 11, at least one camera 33 configured to capture the objects 1a and / or robots in the room.

[0057] The interaction procedure 1 may comprise at least one marker 4 configured to be detected by a sensor 23 at a particular point (e.g., between two rooms 11, objects within a room (and / or between zones of the same room)) that identifies the path of the robot 2 by detecting the passage of the robot 2. The interaction procedure 1 may comprise a marker 4 associated with each passage section 12, such that the sensor 23 detects the passage through that passage section 12.

[0058] Alternatively, or in addition, it may comprise at least one marker 4 associated with each room 11 to enable the sensor 23 to detect the entry of the robot 2 into that room 11 .

[0059] Alternatively, or in addition to one or more markers 4, the robot 2 can detect its own path thanks to sensors 23 configured to acquire and thus identify elements (e.g., furniture, doors, or objects 1 a) along the path. The robot then detects its own path according to its position relative to one or more elements detected by said sensors 23.

[0060] The process 1 may comprise a robot 2 control computer 5. Optionally, it comprises several robots 2 and a computer 5 controlling said robots 2.

[0061] The computer 5 may be configured to control the end effector 21 (and preferably the driver 22) according to the sensors 23 (i.e., one or more of the parameters obtained therefrom) and instructions described below.

[0062] It may be in data connection with the robot 2, in particular with the board 25 and / or the sensor 23.

[0063] The computer 5 may be in data connection with the instruction block 3 .

[0064] The computer 5 may be configured to divide the action into successive scenes according to the scene end commands, in particular it may divide the action into a scene for each scene end command.

[0065] The scene end command can be given automatically and / or manually.

[0066] In the case of an automatic scene end command, the computer 5 may be equipped with a clock configured to measure elapsed time while performing real / virtual actions (described below), and then issue a scene end command at an elapsed time substantially equal to the duration of the scene.

[0067] In the case of a manual scene end command, computer 5 may provide the scene end command upon detecting a command action performed by an operator, such as grabbing a position for at least a threshold time, via block 3. Alternatively, block 3 may comprise a signal unit (such as a button) configured to be activated when the operator wishes to send a scene end command.

[0068] Alternatively, or in addition, an automatic scene exit command may be given by detection of a marker 4 by the robot 2 or by identification of an element of the surrounding environment 10 by the sensor 23 .

[0069] The computer may include an action database, as described below.

[0070] The interaction procedure 1 may include at least one learning phase 6, in which the instruction block 3 defines a virtual action in which the robot 2 moves the end effector 21, the computer 5 commands the robot 2 to perform a real action according to the virtual action, and the sensor 23 acquires one or more parameters while performing the real action.

[0071] The virtual action comprises one or more instructions that identify the movements to be learned by the robot 2 according to the instructions received from the instruction block 3 during the learning phase.

[0072] A real action comprises one or more movements performed by the robot 2 according to the instructions of a virtual action. The movements of a real action are performed according to the instructions of a succession of virtual actions. The real action may therefore be similar to, and optionally the same as, the virtual action.

[0073] The real action may be simulated only (without movement of the robot 2 and / or manipulation of one or more objects 1a) and / or real (with movement of the robot 2 and / or manipulation of one or more objects 1a), preferably real.

[0074] Preferably, the interaction procedure 1 comprises a number of learning phases 6. Thus, the block 3 may define a number of virtual actions, which preferably differ at least in part from each other, and thus the robot 2 performs a number of real actions.

[0075] Each learning phase 6 may comprise a simulation sub-phase 61 in which the instruction blocks 3 define virtual actions by which the robot 2 moves the end effector 21 .

[0076] In the simulation sub-phase 61, the command block 3 defines virtual actions and sends to the robot 2 the commands necessary to perform the real actions corresponding to the virtual actions.

[0077] In some cases, the virtual action may comprise instructions for the robot to pass between two or more rooms 11 , suitably via at least one passage 12 .

[0078] Optionally, one or more scene end commands may be given in the simulation sub-phase 61 .

[0079] The learning phase 6 may comprise an iteration sub-phase 62 in which the computer 5 commands the robot 2 to perform one or more movements that define a real action according to the virtual action.

[0080] In the iteration sub-phase 62, the computer 5 commands the robot 2 to perform one or more movements that define the real action according to the instructions of the virtual action.

[0081] Preferably, in sub-phase 62, the computer 5 commands the robot 2 to substantially repeat the action performed in the virtual action, so that the real action is substantially the same as the virtual action.

[0082] In the iteration sub-phase 62, provided with a virtual action, the robot 2 is moved from one room 11 to another via at least one passage 12.

[0083] The learning phase 6 may comprise an acquisition sub-phase 63 in which the sensor 23 acquires one or more parameters over an iteration sub-phase 62 .

[0084] In sub-phase 63, the sensor 23 may acquire other ambient environment parameters (e.g., brightness, pressure, and / or temperature of the ambient environment 10) and / or one or more interaction parameters, such as the grip of the end effector 21 and / or the force applied thereto.

[0085] Preferably, in the acquisition subphase 63, the sensor 23 can acquire at least the ambient environment parameters (e.g., shape and / or color) of the object 1a to enable the robot 2 to detect the presence of the object 1a in the ambient environment 10.

[0086] Optionally, the sensor 23 can acquire at least one marker 4 to enable the computer to identify the path to be taken. Alternatively, or in addition to the one or more markers 4, the sensor 23 can acquire one or more elements of the surrounding environment 10, and the robot 2 plots its path according to its position relative to the one or more elements detected by the sensor 23.

[0087] In the acquisition subphase 63, the sensors 23 are able to acquire the movements of the robot 2 which make it possible to determine the instructions necessary to define the movements that the robot 2 must subsequently execute to repeat said movements.

[0088] The acquisition sub-phase 63 is completed by sending to the computer 5 the acquisition parameters, one or more scene end commands, and inscriptions regarding the movements of the robot 2.

[0089] Sub-phases 61, 62 and 63 can be performed substantially simultaneously.

[0090] The interaction procedure 1 may comprise an analysis phase 7 in which each of the actual actions performed in the learning phase 6 is divided into successive scenes.

[0091] In the analysis phase 7, the computer 5 can divide each real action into a start scene, an end scene and preferably at least one scene between the start and end scene, suitably according to at least one scene end command.

[0092] The computer 5 can associate with each scene one or more instructions for the robot 2 defining the actions to be performed by the robot, said instructions being defined according to the movements of the robot 2 detected in the acquisition sub-phase 63.

[0093] In some cases, the first scene may not be associated with an instruction.

[0094] The computer 5 may associate each scene with at least one value corresponding to a parameter (environmental and / or interaction) that is obtained during the execution of the scene in the learning phase 6, in particular at the beginning of the scene, i.e. immediately after receiving the scene end command.

[0095] In this document, the term "values" is identified as parameters recorded by sensors 23 in the learning phase 6 and stored in the action database below as appropriate.

[0096] Preferably, the computer 5 associates with each scene a value for each of the parameters obtained in the learning phase 6 .

[0097] The interaction process 1 may include a reprocessing phase 8 in which the computer 5 creates an action database by associating actual actions having substantially the same at least one value with each other that are associated with an initial scene to define a composite action comprising a single initial scene and multiple final scenes.

[0098] Preferably, actual actions with initial scenes having the same value, and optionally the same instruction, are associated with each other.

[0099] The number of final scenes may be substantially equal to the number of real-life actions grouped together.

[0100] If there are one or more intermediate scenes, two or more of these may be merged only if they have substantially the same order and optionally the same values.

[0101] In summary, a composite action has a single initial scene from which several final scenes branch off. If present, intervening scenes can identify one or more branches that connect each final scene to the initial scene.

[0102] In the action database, each scene can be associated with one or more commands for the robot 2.

[0103] Each scene of the synthetic action can be associated with at least one value, i.e., one or more parameters acquired by the sensor 23 during the execution of the scene in each learning phase 6 of the real action that is integrated into that synthetic action. In particular, for each parameter detected by the same sensor during the execution of the scene in the various learning phases 6, the scene is associated with only one value if such parameters are approximately equal to each other; on the other hand, if such parameters are different from each other, the scene is associated with several values, each corresponding to a parameter detected by the sensor in the learning phase 6.

[0104] Each value of a scene is linked to at least one scene that follows it, making it possible to identify the execution order of the scenes. In this document, the expression "next scene" is identified as the scene that is adjacent to or immediately follows a given scene.

[0105] In the case of different values ​​derived from parameters detected by the same sensor, each of these values ​​may be linked to only a part of the subsequent scene. In particular, at least one scene upstream of a branch between alternative scenes of a composite action may be associated with a value, each of which is linked to only a part (in particular only one) of the subsequent alternative scenes, allowing the robot 2 to identify which of the alternative scenes to execute based on that value.

[0106] The interaction procedure 1 may comprise an execution phase 9, in which the robot 2 executes, preferably in automatic mode, one of the actions of the action database.

[0107] In detail, in the execution phase 9, the sensor 23 determines a first factor by detecting the at least one parameter, and the computer 5 selects a composite action in the action database having an initial value substantially equal to the first factor and commands the robot 2 to perform an action according to a scene following the first scene of the identified composite action; the computer 5 commands the robot 2 to perform a next action to be performed from among the next scenes associated with the composite action selected in the action database by comparing at least one value associated with one or more of the previous scenes of the next one (and therefore still to be performed) with at least one factor obtained by detecting at least one parameter in one or more of the previous actions of the next one.

[0108] In this document, the term "factors" is identified as parameters recorded by sensors 23 during the execution phase 9.

[0109] The execution phase 9 may comprise an acquisition sub-phase 91 in which the sensor 23 determines at least one initial factor by detecting said at least one parameter.

[0110] The first factor is obtained at the start of the robot 2, i.e. before it performs any action.

[0111] Preferably, in the acquisition sub-phase 91, the sensor 23 acquires all parameters and then determines an initial factor for each of these.

[0112] During the acquisition sub-phase 91, the robot 2 may be substantially stationary.

[0113] The execution phase 9 may comprise a sub-phase 92 of identifying the composite action to be executed.

[0114] In an identification sub-phase 92, the computer 5 selects composite actions in the action database that have initial values ​​substantially equal to the first factors in question. In particular, the computer 5 selects composite actions in the action database that have initial values ​​substantially equal to all of the first factors defined in sub-phase 91.

[0115] Having identified the composite action, a deployment sub-phase 93 is completed by the computer 3, which commands the robot 2 to execute a scene subsequent to the initial one. The execution phase 9 may comprise at least one deployment sub-phase 93, in which the robot 2 performs actions according to the identified scene, in particular the instructions associated therewith.

[0116] The execution phase 9 may therefore comprise at least one selection sub-phase 94, in which the computer 5 identifies the next scene to be executed among the scenes of the composed action and then commands the robot 2 to perform an action in accordance with that next scene.

[0117] In this sub-phase 94, the computer 5 identifies the next scene from among the next scenes to be executed in the following unfolding sub-phase 93 as appropriate.

[0118] This selection may be performed by comparing at least one value of at least one scene prior to the next scene with one or more of the factors defined in the previously performed action, and the next scene to be performed may be one that is linked to one or more values ​​of at least one previous scene that are substantially equal to one or more factors corresponding to the parameters detected in the previous action.

[0119] This value-factor comparison can be performed for all actions and scenes, up to the earliest ones.

[0120] Once the next scene is identified, the computer 5 commands the robot 2 to perform actions according to the next scene, and then to perform a new deployment sub-phase 93 and a new selection sub-phase 94 .

[0121] The execution phase 9 ends when the last scene is selected as the next scene in the selection sub-phase 94, and then the last action is executed according to that last scene in the next deployment sub-phase 93.

[0122] To clarify the explanation of the procedure, an exemplary application of the procedure of Interaction 1 is given.

[0123] Initially, the procedure comprises the execution of two learning phases 6, namely a first phase 6 comprising a first sub-phase of simulation 61, iteration 62 and acquisition 63, and a second learning phase 6 comprising a second sub-phase of simulation 61, iteration 62 and acquisition 63.

[0124] In the first simulation sub-phase 61, the operator simulates a first virtual action, including locating the glass, picking up the empty glass from the first room 12, reaching the second room 12, filling the glass with the pitcher, and placing the glass on the tray, via instruction block 3. At the end of each part of the virtual action, the operator gives an end scene command that defines the five scenes in this example.

[0125] Simultaneously with the first sub-phase 61, there is a first repetition sub-phase 62 in which the robot 2 repeats the aforementioned actions, guided by the computer 5 according to signals arriving from block 3; and a first acquisition sub-phase 63 in which the sensor 23 acquires, preferably consecutively, glass identification parameters; glass pick-up instructions and the weight of the empty glass; room change instructions 12 (identification of one or more elements and / or one or more markers 4); mug identification parameters and glass filling instructions; and final tray identification parameters (target placement location 1a) and glass placement instructions on the tray.

[0126] Once the first learning phase 6 is completed, the procedure provides for a second learning phase 6 .

[0127] In a second simulation sub-phase 61, the operator simulates a second virtual action consisting of locating the glass, picking up the full glass from the first room 12, reaching the second room 12, and placing the glass on a tray, by means of instruction block 3. At the end of each part of this second virtual action, the operator gives an end scene command that defines this second exemplary four scene.

[0128] Simultaneously with the second sub-phase 61, there is a second repetition sub-phase 62 in which the robot 2, guided by the computer 5 according to signals arriving from block 3, repeats the aforementioned actions; and a second acquisition sub-phase 63 in which the sensor 23 acquires, preferably in succession, glass identification parameters; glass pick-up commands and the weight of the full glass; room change commands 12; and tray identification parameters and glass placement commands on the tray.

[0129] After completing all phases 6, the interaction procedure 1 provides an analysis phase 7 in which the two actions are divided into scenes according to the scene end command with the associated instructions and parameters obtained for each scene.

[0130] The first real-world action is subdivided into an initial scene of locating the glass, a first intermediate scene of picking up the empty glass from the first room 12, a second intermediate scene of reaching the second room 12, a third intermediate scene of filling the glass with the pitcher, and a final scene of placing the glass on the tray.

[0131] The second real-life action is subdivided into an initial scene of locating the glass, a first intermediate scene of picking up the full glass from the first room 12, a second intermediate scene of reaching the second room 12, and a final scene of placing the glass on the tray.

[0132] At this point there is a reprocessing phase 8 of the scene, followed by the creation of the action database.

[0133] In this example, since the initial scenes are the same, the compositing action is defined for the branches of the two final scenes, specifically the end of the second intermediate scene.

[0134] To be precise, the composite action provides a first scene associated with the value "identify the glass", a first intermediate scene associated with a pick-up instruction and two values ​​(weight of the full glass and weight of the empty glass), and a second intermediate scene with room 12 modification instructions (detection of one or more markers 4 and / or one or more elements of the surrounding environment 10). The composite action therefore provides branches (i.e. the possibility of performing different actions): in the first branch of the branching, the composite action comprises a third scene associated with filling the glass and a final scene of placing the glass on a tray; in the second branch, the composite action comprises only the final scene of placing the glass on a tray.

[0135] The first branch is linked to the weight value of the empty glass in the first intermediate scene, and the second branch is linked to the weight value of the full glass in the first intermediate scene.

[0136] At this point, Execution Phase 9 can occur.

[0137] In the acquisition sub-phase 91, the sensor 23 detects / identifies the glasses and thus determines the corresponding initial factors.

[0138] In the identification sub-phase 92, the computer 5 searches the action database for a composite action having a first scene with an initial value substantially equal to the first factor, and commands the execution of the scene following the first scene, i.e., the first intermediate scene. Note that sub-phases 91 and 92 can only be terminated after the detection of a first factor corresponding to an initial value present in the action database.

[0139] Next there is an unfolding sub-phase 93, in which, according to the first intermediate scene, the robot 2 performs an action of picking up a glass (for example, full), and the sensor 23 obtains a factor relative to the weight of the glass.

[0140] At this point, there is a selection sub-phase 94 in which the computer 5 searches among the scenes subsequent to the first intermediate scene (i.e., one or more second intermediate scenes) for one related to the acquisition factor. In this case, since only the second intermediate scene exists, the computer 5 commands execution.

[0141] The execution phase 9 then provides a new sub-phase of execution 93, in which the robot 2 performs actions according to the second intermediate scene. It then moves from the first room 12 to the second room 12, and the sensors 23 acquire acquisition factors corresponding to the markers 4 and / or detected elements.

[0142] Since the last scene has not been executed, a new selection sub-phase 94 occurs in which the computer 5 searches among the scenes following the second intermediate scene for one linked to the acquisition factor.

[0143] Since there are two possible successive scenes, the computer 3 looks among the factors obtained in the previous operations, and those of the previous scenes (i.e., the second intermediate scene, the first intermediate scene, and the first scene) have values ​​corresponding to the factors obtained in the corresponding sub-phase 93. In particular, by identifying the detection of the "full glass" factor in the first sub-phase, it detects that the corresponding value is associated with the last scene of the second case, recognizes the last scene of the second case as the next scene, and commands its execution.

[0144] The execution phase 9 includes a new sub-phase 93 for executing an action according to the last scene. Specifically, the robot 2 searches for a tray in the room 12, identifies it, and places the glasses on the tray and returns it.

[0145] At the end of this unfolding sub-phase 93, phase 9 concludes that the last scene of the identified composite action has been executed.

[0146] The interaction procedure 1 according to the invention achieves important advantages.

[0147] In fact, procedure 1 allows to command robot 2 in a practical and fast way and, above all, to quickly update the specific action database.

[0148] Another advantage, with regard to the creation and use of the action database, lies in the fact that the interaction procedure 1 allows the management of multiple unique actions without complex and expensive structures and without errors in identifying the correct action to be performed.

[0149] This advantage also translates into the possibility of controlling several robots 2 simultaneously in an accurate, simple and fast way.

[0150] Another advantage is the simplicity of commands given to the robot 2.

[0151] Another advantage is therefore identified in the high flexibility of the interaction procedure 1, in fact, that the specific use of sensors 23 allows the robot 2 to adapt to changes in the movements to be performed and / or in the objects 1a with which it interacts, so that continuous reprogramming is not necessary. For example, the robot 2 is in fact able to identify the object 1a regardless of its position and / or location in the surrounding environment 10, and can quickly adapt to any operating state and therefore perform the fastest movements to locate the object 1a.

[0152] The present invention is subject to modification within the scope of the inventive concept as defined by the claims.

[0153] For example, at least in the case of interaction process 1, the card 24 and the computer 5 may match.

[0154] Furthermore, when selecting the next scene in the selection sub-phase 63, the computer can select as the next scene the one associated with the closest value to the factor obtained in at least one previous operation.

[0155] In some cases, the instructions may only identify the final position of the object 1 a and / or the robot 2. The robot 2 then automatically determines the movements to be performed to reach this final position based on parameters obtained by the sensors 23. For example, the instructions may identify the position of the object 1 a on a location, and then the robot 2 determines for itself which movements to perform to place the object 1 a at that location.

[0156] In some cases, the sensor 23 may comprise a camera and the display screen 31 may display the camera image.

[0157] In such an embodiment, all details can be replaced with equivalent elements and materials, and shapes and sizes can be arbitrary.

Claims

1. A process of interaction with an object, At least one robot, said at least one robot comprising: an end effector that interacts with at least one of the objects; a driver for the end effector; a sensor for acquiring at least one parameter, the at least one parameter including at least one of an ambient environment parameter and a parameter of the robot's interaction with the object; a computer for controlling the robot in data communication with the sensor; an instruction block configured to instruct the robot to move the end effector; a plurality of learning phases; and in each of said plurality of learning phases, the instruction block defines a virtual action that the robot performs to move the end effector; The computer commands the robot to perform a real action according to the virtual action; and the sensor acquires the at least one parameter during the performance of the real action; an analysis phase, in which the computer: Each of the real actions is divided into a first scene and a last scene; associating with each of the scenes at least one value corresponding to the at least one parameter acquired in the scene; a reprocessing phase, in which the computer creates an action database associating the real actions containing substantially the same initial values ​​to define a composite action including a single initial scene and multiple final scenes; the action database associating at least one of the values ​​with each of the scenes; The execution phase, in which the sensor determines at least one first factor by detecting the at least one parameter; The computer selects the composite action in the action database that includes the initial value substantially equal to the first factor, and commands the robot to perform an operation substantially equal to a scene subsequent to the first scene of the composite action; and The computer causes the robot to the at least one value associated with the at least one scene prior to the next scene in the action database; the at least one factor obtained by detecting the at least one parameter of the subsequent operation before the subsequent operation; commanding the next action to be performed between the scenes associated with the composite action in the action database selected by comparing; Equipped with The execution phase includes: an acquisition sub-phase in which the sensor determines the first factor by detecting the at least one parameter; an identification subphase, in which the computer selects in a database action that the initial value is substantially equal to the first factor and commands the robot to perform at least one scene subsequent to the first scene; at least one development sub-phase in which the robot executes an action according to the first scene of the composite action selected in the identification sub-phase, and determines a factor by obtaining at least one parameter; at least one selection sub-phase in which the computer identifies the next scene to be executed between the scenes of the composite action and commands the robot to perform an action according to the next scene; and the computer identifies the next scene by comparing the at least one value associated with the at least one scene before the next scene with the factor obtained in the at least one previous action. An interaction process comprising:

2. The interaction process of claim 1, wherein the sensor is configured to acquire the movement of the robot by defining instructions necessary to define the movement of the robot, and wherein in the reprocessing phase the computer associates the instructions with at least one of the scenes in the database action.

3. 3. An interactive process according to claim 1 or 2, wherein the first scene is devoid of instructions.

4. 4. The interaction process according to claim 1, wherein the execution phase ends when one of the last scenes is selected as the next scene and then the robot is commanded to perform a final action according to the last scene.

5. 5. The interaction process according to claim 1, wherein in each of the plurality of learning phases, at least one scene end command is defined; and in the analysis phase, the computer subdivides each actual action into the scenes according to the at least one scene end command.

6. The interactive process of claim 5 , wherein in each of said plurality of learning phases, said at least one scene end command is given automatically.

7. 7. The interactive process of claim 6, wherein the computer comprises a clock configured to measure the passage of time; and in each of the plurality of learning phases, the computer defines the at least one scene end command when a time substantially equal to a scene duration has elapsed.

8. 8. An interaction process according to any one of claims 5 to 7, wherein in each of said plurality of learning phases, said at least one scene end command is given manually.

9. The interactive process of claim 8 , wherein the instruction block comprises a signaling portion configured to transmit the scene end command.

Images