Behavioral programming method for robots in a robotic network architecture

Abstract

The invention relates to a method for behaviorally programming a robot (2) belonging to a robotic network architecture (1), using electronic tags (3). The robotic network architecture (1) comprises at least one robot (2) and at least one element (4). The electronic tags (3) belong to one of the following types: identification and action. The identification type includes subtypes, notably location, object, and person. The method comprises: associating (E1) an identification-type electronic tag (3) with an element (4) of the robotic network (1); coupling (E2) an identification-type electronic tag (3) with an action-type electronic tag (3); and activating (E3) an electronic tag (3) to trigger the robot (2) to perform an action (A) with respect to the element (4) associated with said identification-type electronic tag (3). Abstract: Figure 1

Description

Title of the invention: Method for behavioral programming of a robot in a robotic network architecture Technical field

[0001] The present invention relates to the field of robots, in particular humanoid robots and more specifically aims at a method of behavioral programming of such robots, in a robotic network architecture, by means of electronic labels, also referred to as "tags" or "runes" in the remainder of this document. STATE OF THE ART

[0002] Known from the state of the art is a humanoid robot evolving in a robotic network architecture, in particular a robot adopting a behavior predetermined by a computer program that the robot executes, to achieve a given result.

[0003] The structural characteristics of the robot are in accordance with the behavioral programming conferred on it, itself a function of the task that the robot must carry out. It is then common to “specialize” each robot in the robotic network to obtain the intended result with maximum efficiency. In certain cases, it may be necessary to adapt the environment of the robotic network in which the robot evolves, for example the objects with which the robot must interact according to its programming.

[0004] To overcome the structural constraints of the robot and in particular to avoid a “specialization” of the robots within the network, the applicant has developed, as described in the patent application whose registration number is FR2212216, a robotic network architecture comprising a set of labels, in particular electronic labels, and at least one robot, in particular comprising at least one predefined gripper. The electronic labels are configured to be connected to an element of the robotic network and contain information identifying the element, as well as information allowing the location of the element and / or other information allowing the robot to determine the behavior to adopt with respect to the element.Thanks to the information contained in the electronic label, the robot can in particular determine whether it has the possibility, in particular the capacity and / or the right, to grasp and / or manipulate and / or move, in particular by means of the predefined gripper, said element. The robot is configured to interact with any type of element comprising at least one electronic label and said robot comprises . means of interpreting said identification information, as well as other information possibly contained in the electronic label.

[0005] As is known, the behavior of such a robot can be modified when it is desired to modify the action performed by the robot, in other words the result obtained, for example by directly modifying the computer program executed by said robot. Such modifications of behavior then require specific programming skills, which may exceed the skills of a user or a group of users of the robotic network architecture.

[0006] In fact, the designer or supplier of the robotic network architecture to the user or group of users is generally the only one with the skills or means required to modify the behavior of one or more robots within the robotic network architecture.

[0007] Such an operating model presents many significant disadvantages for the user. On the one hand, the latter is highly dependent on its supplier, particularly when the robot's behavior no longer corresponds to the need and the programming must be modified.

[0008] On the other hand, the development of new behavioral programming can be long and financially expensive, without the end users of the robot necessarily being involved in the development of this new behavioral programming. The risk of mismatch between the behavior of the robot and the need is then significantly increased.

[0009] The present invention therefore aims to solve the aforementioned problems by proposing a simplified and accessible method for programming the behavior of a robot. PRESENTATION OF THE INVENTION

[0010] More specifically, the subject of the invention is a method for behavioral programming of a robot belonging to a robotic network architecture, from electronic labels, the robotic network architecture comprising at least one robot and at least one element, the electronic labels belonging to a type among the identification and action types, the identification type comprising subtypes, in particular location, object and person, the method comprising: • the association of at least one electronic identification-type label with at least one element of the robotic network; • the coupling of at least one identification-type electronic label to at least one action-type electronic label; • the activation of at least one electronic label to cause the robot to perform an action with respect to the element associated with said identification-type electronic tag coupled to said action-type electronic tag.

[0011] Thanks to such a combination of characteristics, the programming of a given behavior by a user or a group of users of the robotic network is significantly simplified and does not require a computer development phase. Also, the user can then create in situ a new behavior of a given robot, in other words a new rule intended to generate a desired behavior of the robot in defined circumstances, without requiring major structural or software adaptation of the robot or the robot's environment.

[0012] The element of the robotic network with which the identification-type electronic tag is associated also comprises a subtype, in particular location, object or person. An electronic tag of a given subtype is preferably associated with an element of the same subtype. An electronic tag of the person subtype is associated with an element of the person subtype, such as a network user, with which the robot can interact. An electronic tag of the location subtype is associated with a location subtype element to which the robot can navigate. A tag of the object subtype is associated with an object subtype element that the robot can manipulate. The behavioral programming obtained through the method using these subtypes associated with an electronic tag can make it possible to carry out a large number of types of social logistics action.

[0013] Identification type electronic tags can be used to make information available to simplify the action of the robot and to accelerate its reaction time and to reduce the risk of reducing the risk of error, by reinforcing the principle of simplification of the programming of the invention. This simplification also has the advantage of reducing energy consumption across the entire network, making it possible to reduce the operating cost of such an infrastructure, as well as improving the autonomy of the robot. For example, an electronic tag of object subtype can include information on the left or right hand of the robot which must manipulate the object subtype element associated with said tag.

[0014] Advantageously, the coupling step corresponds to the definition or modification of a rule configured so that the robot interprets the activation of at least one electronic label to carry out a predefined action with respect to the element associated with said identification-type electronic label coupled to said action-type electronic label. In such a configuration, the possibilities of action can be widely extended.

[0015] An electronic label, in particular of the action type, can be associated with a reference number, for example a positive integer. Therefore, the labels identification-type electronics are coupled to one or more action-type labels to generate at least one or more complex actions with respect to one or more elements.

[0016] Advantageously, the method comprises a step of decoupling at least one electronic label of identification type from at least one label of action type.

[0017] Advantageously, the association and coupling steps are carried out by means of instructions in natural language using a suitable interpretation system or by activation of a visual or tactile command, by means of a human-machine interface. The term "natural language" means a conversational language different from a computer programming language. In this configuration, any user can carry out an association or coupling step, for example by means of an oral or written instruction when the robot is nearby or remotely by means of a terminal connected to the electronic tags of the network.Furthermore, in natural language, when an electronic label has been associated with an element of the robotic network, said label is designated by extrapolation when an instruction mentions said element of the associated network, simplifying the formulation of an action or a rule for a user. Advantageously, the activation step can also be carried out via instructions in natural language.

[0018] Advantageously, the action-type electronic tag comprises a plurality of activation modes, among which the electronic tag is: pressed; shaken; moved; or upon arrival in proximity to the corresponding electronic tag, i.e. arrival at a distance less than a predefined activation distance, of an element associated with another identification-type electronic tag.

[0019] Advantageously, each given activation method of the action-type electronic tag corresponds to a predetermined action. In such a configuration, a single electronic tag is necessary to be associated with an element of the network and generate an action of the robot upon its activation.

[0020] Advantageously, an electronic label can be both of the identification type and of the action type.

[0021] Advantageously, the robot carries at least one electronic label of the identification type and / or of the action type.

[0022] Advantageously, an element of the robotic network is associated with at least two electronic labels, including at least one electronic label of the identification type.

[0023] Advantageously, an element of the network, in particular a user of the robotic network, is associated with an identification-type electronic label, the coupling of said user's electronic label to an action-type electronic label consisting of the creation of a rule, the activation of said electronic action-type label generating at least one action among: • the manipulation by the robot of an element associated with an electronic label of identification type and object subtype; • the robot's navigation within the robotic network from a first point to a second point; • the response to a request.

[0024] According to another aspect of the invention, it relates to a system for behavioral programming of a robot in a robotic network, configured to implement the method as described above. Such a system may comprise at least one computer and at least one human-machine interface. PRESENTATION OF FIGURES

[0025] The invention will be better understood on reading the following description, given solely by way of example, and referring to the appended drawings given by way of non-limiting examples, in which identical references are given to similar objects and in which:

[0026] [Fig.l] is a schematic representation of a flowchart of the behavioral programming method according to a first embodiment of the invention;

[0027] [Fig.2] is a schematic representation of a flowchart of the behavioral programming method according to a second embodiment of the invention;

[0028] [Fig. 3] is a schematic representation of a flowchart of the behavioral programming method according to a third embodiment of the invention;

[0029] [Fig.4] is a schematic representation of a robotic network according to a first use case of the invention;

[0030] [Fig.5] is a schematic representation of a robotic network according to a second use case of the invention;

[0031] [Fig.6] is a schematic representation of a robotic network according to a third use case of the invention;

[0032] [Fig.7] is a schematic representation of a robotic network according to a fourth use case of the invention;

[0033] [Fig.8] is a schematic representation of a first example of robot behavior obtained by the behavioral programming method of the invention;

[0034] [Fig.9] is a schematic representation of a second example of robot behavior obtained by the behavioral programming method of the invention;

[0035] [Fig. 10] is a schematic representation of a third example of robot behavior obtained by the behavioral programming method of the invention;

[0036] [Fig. 11] is a schematic representation of a fourth example of robot behavior obtained by the behavioral programming method of the invention;

[0037] [Fig. 12] is a schematic representation of a fifth example of robot behavior obtained by the behavioral programming method of the invention;

[0038] [Fig. 13] is a schematic representation of a sixth example of robot behavior obtained by the behavioral programming method of the invention;

[0039] [Fig. 14] is a schematic representation of a seventh example of robot behavior obtained by the behavioral programming method of the invention;

[0040] [Fig. 15] is a schematic representation of an eighth example of robot behavior obtained by the behavioral programming method of the invention.

[0041] It should be noted that the figures set out the invention in detail to enable the invention to be implemented; although not limiting, said figures serve in particular to better define the invention where appropriate. DETAILED DESCRIPTION OF THE INVENTION

[0042] The invention relates to a method for behavioral programming of a robot 2 belonging to a robotic network architecture 1, the method being implemented by a behavioral programming system. The behavioral programming of the robot 2 is carried out using electronic tags 3. The electronic tags 3 each comprise a capsule, a transmitting antenna and a battery, the transmitting antenna and the battery being housed in said capsule. The robot 2 comprises in particular a receiving antenna configured to communicate with said transmitting antenna of the electronic tags 3.

[0043] The electronic labels 3 may belong to one of the types identification, parameter and action. An electronic label 3 of identification type also belongs to a subtype, in particular location, object or person.

[0044] The behavioral programming method includes in particular the following steps, as illustrated in [Fig.l]: • the association El of at least one electronic label 3 of identification type with at least one element 4 of the robotic network 1; • the coupling E2 of at least one electronic label 3 of identification type to at least one electronic label 3 of action type; • the activation E3 of at least one electronic label 3 to cause the robot 2 to carry out an action A with respect to the element 4 associated with said electronic label 3 of identification type coupled with said electronic label 3 of action type.

[0045] Thanks to such a combination of characteristics, the programming of a given behavior by a user or a group of users of the robotic network 1 is significantly simplified and does not require a specific computer development phase for each behavior or each type of behavior. Also, the user can then create in situ a new behavior of a given robot 2, in other words a new rule intended to generate a desired behavior of the robot 2 in defined circumstances, without requiring major structural or software adaptation of the robot 2 or of the robot's environment.

[0046] The element 4 of the robotic network 1 with which the electronic label 3 of the identification type is associated also comprises a subtype, in particular location, object or person. An electronic label 3 of a given subtype is preferentially associated with an element of the same subtype. An electronic label 3 of the person subtype is associated with an element 4 of the person subtype, such as a user of the network 1, with which the robot 2 can interact. An electronic label 3 of the location subtype is associated with an element of the location subtype towards which the robot 2 can navigate. An electronic label 3 of the object subtype is associated with an object subtype element that the robot can manipulate.When element 4 is a person in the robotic network 1, the latter may or may not have the authorization to program the behavior of the robot, for example with an electronic tag type 3 of identification type and person subtype having corresponding rights.

[0047] The “parameter” type of electronic label 3 includes, in a non-limiting manner, physical quantity sub-type labels, for example a distance, a duration, a frequency, a temperature, etc. The “parameter” type of electronic label 3 may also include, in a non-limiting manner, logical operator sub-type labels, for example “and”, “or”, “not”, or adverbs, for example “never”, “always”, “now”, “after” etc.

[0048] The elements 4 associated with electronic labels 3 of identification type and object subtype, the robot being provided with a gripper so that the robot 2 can manipulate the element 4 whatever its shape. In this case, the electronic labels 3 can in particular be integrated into handles as described in document FR2212228. This gives the robot instantaneous information on the means by which the object can be manipulated, namely the handle. According to one embodiment, the electronic labels 3 associated with elements 4 of object subtype, in particular integrated into handles, further comprise additional information, with the aim of facilitating and making the manipulation of said objects by the robot faster and more efficient.In particular, an electronic label 3 of object subtype may include information on the left or right hand of the robot which must manipulate the element 4 of object subtype associated with said label. This information may be a function of the angle of arrival of the robot relative to the object, when said angle is available. Also, or alternatively, the electronic label 3 may also include information according to which the handle is arranged horizontally, or vertically, or inclined with a certain angle on the element 4 of subtype to be manipulated. Or again, the electronic label 3 may include information according to which there is an obstruction on one side or the other of said element 4 of object subtype.

[0049] According to an embodiment of the programming method, the coupling step E2 corresponds to the definition E21 or the modification E21' of a rule configured so that the robot interprets the activation E3 of at least one electronic label 3, as illustrated in [Fig.2], to carry out a predefined action A with respect to the element associated with said electronic label 3 of identification type coupled to said electronic label 3 of action type. The predefined action A is in particular predefined according to the subtype of the electronic label of identification type coupled to the electronic label of action type 3 as described previously. Such a definition E21 or modification E21' of rule thus corresponds to a programming sequence.

[0050] According to one embodiment of the programming method, the association E1 and coupling E2 steps are carried out by means of instructions E0 in natural language, for example by means of a suitable interpretation system or by visual activation on a touch screen, as illustrated in [Fig.2]. The term "natural language" means a conversational language other than a computer programming language. In this configuration, any user can carry out an association E1 or coupling E2 step, for example by means of an oral or written instruction E0 when the robot is nearby or remotely by means of a terminal connected to the electronic labels of the network. In addition in natural language, when an electronic label 3 has been associated with an element 4 of the robotic network 1, said label 3 is designated by extrapolation when an instruction mentions said element 4 of the associated network, simplifying the formulation of an action A or a rule for a user. For example, if the element 4 associated with the electronic label 3 is a table, in a natural language instruction, the word "table" will be interpreted by the robot 2 as designating the electronic label 3 of identification type and object subtype associated with the element 4 which is a table. The robot 2 preferably executes a natural language processing algorithm based on a large language model, commonly referred to as a "large language model" or "LLM" by those skilled in the art.

[0051] Depending on the situations, it is possible that several behaviors are programmed simultaneously for the same robot 2. In such a configuration, the robot 2 carries out in particular the actions A by following a hierarchy predefined by the users of the network 1, particularly in the case where the behaviors are incompatible with each other.

[0052] The remainder of the description focuses on describing different use cases exploiting the behavioral programming method, in particular according to different configurations of the robotic network 1 as illustrated in Figures 4 to 7.

[0053] [Fig.4] illustrates the robotic network 1 according to a first use case. Here, the robotic network 1 is a professional premises of a company. Each employee of the company is an element 4 of the person subtype of the robotic network 1. The association step E1 is for example carried out upon the arrival of the employee at the premises, the employee keeping the associated electronic label 3 on him at all times. To distribute a parcel received by a receptionist of the company, said receptionist carries out an association E1 of an electronic label 3 of identification type and of object subtype to the parcel element 4 of object subtype. Then, the receptionist carries out the coupling E2 of an electronic label 3 of action type to the electronic label 3 associated with the parcel, for example by stating a rule in natural language. The rule or instruction can for example be stated in the following manner: “Take the parcel to Mr. X”.The receptionist then performs the activation E3 of the electronic label 3 of the coupled action type, either when the coupling E2 is completed and the rule is stated, or by manual activation of the electronic label 3, for example by pressing the capsule of the electronic label 3 of the action type. Other examples of activation E3 of the electronic label 3 are described below. The robot 2 then performs the action corresponding to the stated rule, which can for example be broken down as follows: . • confirm correct understanding of the rule; • handle the package using the gripper; • navigate within the company premises; • deliver the package to its recipient.

[0054] [Fig.5] illustrates the robotic network 1 according to a second use case. Here, the robotic network 1 is a shopping mall, in particular a food court comprising a plurality of restaurants. When a customer orders a food product from one of the restaurants, the server of said restaurant performs the association E1 of an electronic label 3 of identification type and subtype person with the customer element 4 of subtype person. When the food product is ready to be served, the server places it on a tray already provided and associated with an electronic label of identification type and subtype object. Then, the server performs the coupling E2 of an electronic label 3 of action type with the electronic label 3 associated with the customer, for example by stating a rule in natural language: “Take the order to customer X”.The server then performs the activation E3 of the electronic label 3 of the coupled action type, either when the coupling E2 is completed and the rule is stated, or by manual activation of the electronic label 3 as described above. The robot 2 then performs the action A corresponding to the stated rule, which can for example be broken down as follows: . • manipulate the tray using the gripper; • navigate the food court to the associated customer's table; • serve the tray to the associated customer.

[0055] Not exclusively for this use case, the coupled action type electronic label 3 may be the identification type electronic label 3, so that only one capsule is required. An electronic label may advantageously be of the identification type and of the action type.

[0056] [Fig.6] illustrates the robotic network 1 according to a third use case. Here, the robotic network 1 is a hospital. At the time of taking treatment for a patient staying in a room of the hospital, the nurse performs the association E1 of an electronic label 3 of identification type and location subtype with the room element 4 of location subtype.

[0057] The nurse places the treatment on a tray already equipped and associated with an electronic label 3 of identification type and object subtype. Then, the nurse performs the coupling E2 of an electronic label 3 of action type to the electronic label 3 associated with the room, for example by stating a rule in natural language: “Take the treatment to room X and come back to confirm that the treatment has been taken”. The nurse then performs the activation E3 of the coupled action type electronic label 3, either when the coupling E2 is finished and the rule is stated, or by manual activation of the electronic label 3 as described above. The robot 2 then performs action A corresponding to the stated rule, which can for example be broken down as follows: • manipulate the tray using the gripper; • navigate the hospital; • provide treatment to the associated patient; • return to the nurse to confirm that the treatment has been taken.

[0058] Naturally in this use case, it is understood that the nurse is an element 4 of the robotic network 1 of subtype person, and that he has an electronic label 3 of type identification and subtype person, so that the robot 2 can interpret the rule stated during the coupling E2. The electronic label 3 of type action has then been coupled not only to the electronic label 3 of the room, but also to the electronic label 3 associated with the nurse.

[0059] [Fig.7] illustrates the robotic network 1 according to a fourth use case. Here, the robotic network 1 is a hotel. When a customer arrives at the hotel reception, the receptionist performs the association E1 of an electronic label 3 of identification type and person subtype to the customer element 4 of person subtype. Then, the receptionist performs the coupling E2 of an electronic label 3 of action type to the electronic label 3 associated with the customer as well as to an identification type label associated with a room element 4, for example by stating a rule in natural language: “Guide customer X to his room Y”. The server then performs the activation E3 of the coupled action type electronic label 3, either when the coupling E2 is finished and the rule is stated, or by manual activation of the electronic label 3 as described above. The robot 2 then performs the action A corresponding to the stated rule.

[0060] The remainder of the description focuses on describing different behaviors of robot 2, obtained by the behavioral programming method of the invention. Depending on the desired behaviors, an electronic tag 3 can advantageously be both of the identification type and of the action type, which makes it possible to limit the number of electronic tags 3 required, in particular when the association E1 is made with an element 4 of the location subtype.

[0061] [Fig.8] illustrates in particular the navigation of the robot 2 within the network 1, by successive coupling E2 of an electronic tag 3 of action type to electronic tags 3 of identification type and location subtype. In the case of navigation towards a location, the capsules of the electronic tags 3 comprise in particular Bluetooth chips satisfying the 5.1 standard or higher and comprising a radiogoniometry function making it possible to determine an angle of arrival relative to another electronic tag 3 whose capsule comprises a Bluetooth chip satisfying the 5.1 standard or higher and comprising a function compatible radio direction finding. E3 activation is for example carried out when the action type electronic tag 3 is at a distance less than a predefined activation distance from a coupled identification type electronic tag 3.

[0062] [Fig.9] illustrates in particular another action A of the robot 2 within the network 1, by coupling E2 of an electronic tag 3 of the action type to an electronic tag 3 of the identification type. Here, the action generated during activation E3 is a restriction, that is to say the definition of a path in which the robot 2 cannot circulate, for example so as not to enter a stairwell which could cause the robot 2 to fall, or the definition of a particular zone in which the presence of the robot is prohibited.

[0063] [Fig. 10] illustrates in particular another action A of the robot 2 within the network 1, by coupling E2 of an electronic tag 3 of action type to two electronic tags 3 of identification type. Here, the action generated during the action E3 is a manipulation, that is to say the grasping of an element 4 of object subtype, in particular a tray comprising two grippers containing the electronic tags of identification type. Advantageously, the two electronic tags 3 of identification type can include information on the left or right hand of the robot 2 which must manipulate the element of object subtype associated with said electronic tag 3.

[0064] [Fig. 11] illustrates in particular another action A of the robot 2 within the network 1, by coupling E2 of an electronic tag 3 of action type to an electronic tag 3 of identification type. Here, the action A generated during activation E3 is a manipulation, in particular after the identification of a reception zone in which the robot 2 must deposit a previously manipulated object as illustrated in [Fig. 10], by locating the electronic tag 3 of said reception zone. The electronic tag 3 of the reception zone of the object can advantageously comprise information on the preferential orientation that the robot 2 must give to the manipulated object when depositing it in the reception zone. In such a configuration, the action A of the robot is significantly accelerated and the risk of manipulation error is reduced, while reinforcing the principle of simplification of the programming of the invention.This simplification also has the advantage of reducing energy consumption across the entire network, reducing the operating cost of such an infrastructure, as well as improving the robot's autonomy.

[0065] [Fig. 12] illustrates in particular another action A of the robot 2 within the network 1, by coupling E2 of an electronic tag 3 of the action type to an electronic tag 3 of the identification type. Here, the coupling E2 has been previously carried out, and the activation E3 is carried out by means of a voice command in natural language of the part of a person corresponding to an element 4 of the network. The action A generated is the execution of a personalized audio track according to the coupled electronic label 3, in particular according to the element 4 to which the electronic label 3 is associated, for example to greet it in return for a greeting.

[0066] In some cases, there may be a hierarchy between several stated rules, as illustrated in [Fig. 13]. Here, the activation E3 of the electronic label 3 of the location subtype associated with the zone and coupled according to a rule formulated for example in the following way: "No noise in this zone"; takes priority over the activation E3 of the electronic label 3 of the person subtype associated with the element 4 person and whose generated action A would be that described in [Fig. 12]. This action is then not carried out.

[0067] [Fig. 14] illustrates in particular another action A of the robot 2 within the network 1, by coupling E2 of an electronic tag 3 of the action type to an electronic tag 3 of the identification type. Here the action A generated during the activation E3 is a navigation, for example following the enunciation of a rule during the coupling E2 with the electronic tag 3 of the person subtype of the associated element 4, in the following manner: “Follow me”.

[0068] [Fig. 15] illustrates in particular another action A of the robot 2 within the network 1, by coupling E2 of an electronic tag 3 of the action type to an electronic tag 3 of the identification type. Here, the action A generated during the activation E3 is a dance of the robot 2. The activation E3 is also carried out when the robot 2 enters the range of the electronic tag 3 of the location subtype.

[0069] A single action-type electronic tag 3 can be activated in several different ways. In particular and advantageously, when an action-type electronic tag 3 is coupled to several identification-type electronic tags 3, each activation method can correspond to a given coupled identification-type electronic tag 3. The electronic tag 3 can for example be pressed, shaken or moved.

[0070] The behavioral programming obtained by means of the method of the invention, in particular using electronic labels of the identification type and having the subtypes described, can make it possible to carry out any type of social logistics action A, from simplistic algorithmic instructions and not requiring software development skills.

[0071] According to another embodiment of the invention, two electronic labels 3 of identification type and / or parameter type are coupled to the same electronic label 3 of action type. In such a configuration, the action A generated then involves the two electronic labels 3 used and the elements 4 of the robotic network 1 associated to form combinations. To be coupled to the same electronic tag 3 of action type, the two electronic tags of identification and / or action type can for example be shaken together in the hands of a user of the robotic network 1.

[0072] Advantageously, the robot 2 initiates a conversational sequence, in particular in natural language, orally or via a human-machine interface (touch screen, etc.), with a user of the robotic network 1 to arrive at the definition E21 or the modification E21' of a rule. The conversational sequence preferably consists of a succession of interactions between a robot 2 and a user of the robotic network 1, in particular by means of a dialogue in natural language. Such a conversational sequence has the advantage of allowing the user to specify the definition E21 or the modification E21' of a rule from a very large list of rules resulting from the combination of several electronic labels.

[0073] In this context, the electronic labels 3 allow the natural language processing algorithms used to converge much more quickly through the conversational sequence towards a rule satisfying the needs of the user, the rule necessarily having to involve the electronic labels 3 of the identification type and / or coupled with the electronic label 3 of the action type. Thus, the natural language processing times of the robot 2 are significantly reduced. Furthermore, such a conversational sequence also allows users of the robotic network who are not trained or who are not experts in software programming to simply formulate or modify rules, the electronic labels 3 making it possible to manipulate the behavioral programming syntax.

[0074] For example, a user having an electronic tag 3 of parameter type and physical quantity subtype “hour”, an electronic tag 3 of identification type and location subtype “room” and an electronic tag of action type generating the action “dance” shakes the three electronic tags 3 together in his hands to create a coupling E2.

[0075] The robot 2 then detects a combination during the E2 coupling, and initiates a conversational sequence in natural language with the user. First, the robot 2 describes to the user its interpretation of the E2 coupling of the three electronic labels 3 described above, to form a rule prototype. Here, the robot 2 has interpreted a rule prototype “dance in the bedroom at a given time”. The user can thus confirm the success of the E2 coupling.

[0076] Starting from the rule prototype, robot 2 can then ask the user for clarification. In the example described, robot 2 asks the user "what time should I go dancing in the bedroom?". The user can then respond, here for example, by answering "nine o'clock in the morning." The dialogue continues until convergence towards the rule initially imagined by the user.

[0077] In other words, when the user initiates a robot programming sequence by means of the electronic labels 3, as described previously, the robot initiates a conversational sequence to dispel any possible ambiguity. The conversational sequence aims to converge towards a certain interpretation validated by the user of the action to be carried out, resulting in the effective behavioral programming of the robot which will execute the rule imagined by the user.

[0078] The robot generates sentences using a generative AI algorithm, in particular based on one or more natural language processing algorithms based on a large language model, of the “LLM” type. The sentences thus generated are vocally synthesized for the user or are displayed on a screen, for example a touch screen.

[0079] The generated sentences are designed to elicit short and clear responses.

[0080] In particular, the generated sentences correspond to behavioral proposals and the user selects, via an oral response or a selection on the screen, the desired behavior to validate it, in other words so as to confirm the rule to be implemented by the robot.

[0081] The rule defined or modified by means of a conversational sequence as described above can advantageously be uploaded to an action-type electronic label 3 configured so that the robot 2 interprets the activation E3 of said electronic label 3 to carry out an action A corresponding to the rule resulting from the conversational sequence.

[0082] In the case where the electronic tag 3 is associated with a user of the robotic network 1, said electronic tag 3, in particular of the identification type, may also be configured to transmit a signal to the user, for example by means of a haptic signal such as a vibration. Such a haptic signal may for example allow the user to determine the success or failure of a step of the behavioral programming method, the signal being able to be differentiated according to said step to correspond either to a coupling E2, to a definition E21 of a rule, to a modification E21' of a rule, or to the activation E3 of an electronic tag 3.

[0083] Advantageously, the robotic network 1 comprises at least one electronic tag 3 of identification type and administrator subtype. The element 4 with which such an electronic tag 3 is associated has specific authorizations within the robotic network 1, for example the possibility of consulting the robot 2 to check the data in its possession, the rules currently known. Such a user thus has the necessary authorization, in particular in the event of a malfunction requiring a higher degree of software development skill.

[0084] It will also be noted that the invention is not limited to the embodiments described above. It will indeed appear to those skilled in the art that various modifications can be made to the embodiments described above, in light of the teaching which has just been disclosed to them.

[0085] In the detailed presentation of the invention which is made above, the terms used must not be interpreted as limiting the invention to the embodiment set out in the present description, but must be interpreted to include all equivalents whose prediction is within the reach of a person skilled in the art by applying his general knowledge to the implementation of the teaching which has just been disclosed to him.

Claims

Claims

1. Method for behavioral programming of a robot (2) belonging to a robotic network architecture (1), from electronic labels (3), the robotic network architecture (1) comprising at least one robot (2) and at least one element (4), the electronic labels (3) belonging to one type among the identification and action types, the identification type comprising subtypes, in particular location, object and person, the method comprising: • the association (El) of at least one electronic label (3) of identification type with at least one element (4) of the robotic network (1); • the coupling (E2) of at least one electronic label (3) of identification type with at least one electronic label (3) of action type;• the activation (E3) of at least one electronic label (3) to cause the robot (2) to perform an action (A) with respect to the element (4) associated with said electronic label (3) of identification type coupled to said electronic label (3) of action type; the method being characterized in that the coupling step (E2) corresponds to the definition (E21) or the modification (E21') of a rule configured so that the robot (2) interprets the activation (E3) of at least one electronic label (3) to perform a predefined action (A) with respect to the element (4) associated with said electronic label (3) of identification type coupled to said electronic label (3) of action type.;

2. Method according to claim 1, in which the association (El) and coupling (E2) steps are carried out by means of instructions in natural language (EO) by means of a suitable interpretation system or by activation of a visual or tactile command, by means of a human-machine interface.

3. A method according to one of the preceding claims, wherein the action-type electronic tag (3) comprises a plurality of activation modes, among which the electronic tag (3) is: pressed; shaken; moved or upon arrival in the vicinity of the corresponding electronic tag (3), i.e. arrived at a distance less than a predefined activation distance, from an element (4) associated with another electronic tag (3) of identification type.

4. Method according to the preceding claim, in which each given activation method of the electronic tag (3) of action type corresponds to a predetermined action (A).

5. Method according to one of the preceding claims, in which the electronic labels (3) can further belong to a parameter type.

6. Method according to one of the preceding claims, in which an electronic label (3) can be of at least two types at a time among the identification type, the parameter type and the action type.

7. Method according to one of the preceding claims, in which an element (4) of the robotic network (1) is associated with at least two electronic labels (3) including at least one electronic label (3) of the identification type.

8. Method according to one of the preceding claims, in which an element (4) of the network (1), in particular a user of the robotic network (1) is associated with an electronic label (3) of identification type, the coupling of said electronic label (3) of the user to an electronic label (3) of action type consisting of the creation of a rule, the activation of said electronic label (3) of action type generating at least one action (A) among: • the manipulation by the robot (2) of an element (4) associated with an electronic label (3) of identification type and of object subtype; • the navigation of the robot (2) within the robotic network (1) from a first point to a second point; • the response to a request.

9. Method according to one of the preceding claims, in which at least two electronic labels of identification type and / or parameter type are coupled to the same electronic label of action type.

10. Method according to one of the dependent claims, in which the robot (2) initializes a conversational sequence in natural language with a user of the robotic network (1) to result in the definition (E21) or modification (E2E) of a rule.

11. Method according to claim 10, in which the conversational sequence consists of a generation of sentences, the generated sentences corresponding to behavior proposals, the user selecting the desired behavior to validate it.

12. Method according to claim 10 or 11, in which the rule defined or modified by means of a conversational sequence is uploaded to an action-type electronic label (3) configured so that the robot (2) interprets the activation (E3) of said electronic label (3) to carry out an action (A) corresponding to the rule resulting from the conversational sequence.

13. System for behavioral programming of a robot (2) in a robotic network (1), configured to implement the method according to any one of the preceding claims.

Images