System for transporting objects employing at least one reusable smart mobile container and a chatbot

Abstract

The invention relates to a system for transporting objects allowing a conversational exchange. According to the invention, such a system comprises at least one reusable smart mobile container for transporting objects (1), said container being configured to capture a first datum originating from an interlocutor, an identifier of the object present in the container, and a chatbot (18) configured to: receive the first datum with the object identifier; perform a contextualised semantic analysis of the first datum in the light of the object identifier; optionally produce a second contextualised datum resulting from the semantic analysis; optionally transmit the second datum to said at least one container; said container (1) further being configured so as to process the received second datum.

Description

[0001] Object transport system implementing at least one reusable intelligent mobile container and a conversational assistant. Field of the invention

[0002] The invention relates to the field of conversational artificial intelligence linked to connected objects, also known under the terminology of the Internet of Things and the acronym loT (from the English “Internet of Things”).

[0003] More specifically, the invention relates to a smart container, intended for the transport of products, parts or objects.

[0004] The invention finds particular application in packaging or delivery packages, used in the field of logistics transport for the delivery of products. . State of the art

[0005] In the field of logistics, it is known to label a package to identify its destination, sort it and direct it on logistics platforms, until its recipient. Such a technique suffers from the need to multiply the number of labels depending on the number of logistics providers involved in the transport. In addition, this elementary technique has many disadvantages related to the absence of the peeled label, legibility problems, and the impossibility of tracking the package between two identification points in logistics platforms. As a result, many packages are lost and some stolen.

[0006] It then became necessary to use packaging equipped with an on-board electronic unit with a screen acting as a label and capable of transmitting position information with a monitoring platform, on the logistics route followed. Finally, to reinforce the security of these packages, means of locking the packages were added, authorizing opening under specific conditions, in order to limit access to the authorized person, very often the final recipient.

[0007] However, these developments in packaging for shipping and transporting an object are not without drawbacks.

[0008] Indeed, as soon as a new type of labeling comes into operation, it is necessary to review and reformulate the processes, consider the nominal operation and the situations of processing errors in the supply chain.

[0009] Another disadvantage of such a technique is that it requires training all professionals in the supply chain in this new, often specific, use.

[0010] Another disadvantage of this technique is that when the recipient receives his package, although he has been able to follow its progress, he has no information relating to its transport. Objectives of the invention

[0011] The invention therefore aims in particular to overcome the drawbacks of the state of the art cited above.

[0012] More specifically, the invention aims to provide an object transport system allowing a conversational exchange between a user and the transport packaging.

[0013] An objective of the invention is also to provide such a technique that is simple to implement by the user. Another objective of the invention is to provide such a technique that allows relevant exchanges with the user.

[0014] An objective of the invention is also to propose such a technique which allows a conversational exchange centered around the sender's profile.

[0015] Another objective of the invention is to provide such a technique which is reliable and safe for the user. . Statement of the invention

[0016] These objectives, as well as others that will appear later, are achieved using an object transport system that allows for a conversational exchange with an interlocutor.

[0017] In the context of the invention, the system comprises at least one reusable intelligent package for transporting objects, also called a container adapted to contain an object and capable of conversing with an interlocutor. The package is configured so as to capture a first piece of data; the system also comprises an object identifier, of the object contained in other words present in the container during the transport of said object.

[0018] The system further comprises a conversational intelligent agent configured to: a) receive the first data coupled with the object identifier contained during transport in the container, the first data being transmitted by the container, the object identifier being further transmitted so as to allow a conversational exchange relating to the transported object; b) perform a semantic analysis of the first data received, contextualized in light of the object identifier coupled with the first data, forming a semantic result; c) produce, optionally, a second contextualized data, resulting from the semantic analysis; d) transmit, optionally, the second data to said at least one container, in response to the first data;

[0019] Furthermore, in this system, the container is configured to process the second received data.

[0020] Thus, in such a system, thanks to these technical characteristics it is possible for a user to become an interlocutor of the package and to initiate a conversation in which the container sends data to an intelligent agent which, after having semantically analyzed them, can respond with the result of its analysis.

[0021] According to a particularly advantageous embodiment of the invention, the intelligent conversational agent is configured to:

[0022] - select a first corpus of initial conditions based on a selection criterion, from a library or set of corpora of initial conditions;

[0023] - use the first corpus of initial conditions when producing the second contextualized data.

[0024] Thanks to these technical features, the volume of input data to the model is significantly smaller, which also has the advantage of faster processing.

[0025] According to an advantageous embodiment of the invention, the selection criterion is a function of the object identifier, and / or the semantic result of the analysis of the first data. Thanks to these technical characteristics, the targeted data sought are targeted with a corpus of initial conditions chosen accordingly. This has the advantage of increasing the relevance of the answers to the questions of the user interlocutor of the smart parcel.

[0026] According to a preferred embodiment of the invention, the first corpus of initial conditions comprises information relating to the transported object, to enable the conversational intelligent agent to maintain a conversation relating to said object.

[0027] Thanks to these technical features, the reusable smart parcel for transporting objects can respond appropriately to the recipient's questions on all subjects relating to the product or object delivered.

[0028] In addition, the selection of a specialized corpus makes it possible to increase the reliability of the responses delivered and limit hallucination phenomena.

[0029] According to a particular embodiment of the invention, the intelligent conversational agent is configured to select at least one second corpus of specialized initial conditions based on the selection criterion, in the library, the first corpus and the at least one second corpus forming a group of corpora.

[0030] According to an advantageous embodiment of the invention, the conversational intelligent agent is configured to:

[0031] - extract a subset of data from the corpus group using a filter based on the selection criterion;

[0032] - use said subset of data in the production of the second contextualized data. Thanks to these technical characteristics, the intelligent conversational agent can adapt to different situations. In particular, the intelligent conversational agent can handle cases where several objects are present in the delivered package. In such a situation, an object identifier is transmitted to it for each object included in the package. In this way, it can, for example, select a corpus of initial conditions per object. However, the multiple selection of corpuses linked to the simultaneous delivery of several objects in the same package should not be understood as a limitation.

[0033] According to a particular embodiment of the invention, the container is configured to

[0034] - store the object identifier of the object present in the container, and

[0035] - attach the stored object identifier, coupled with the first data transmitted to the conversational intelligent agent.

[0036] It is understood that the container is configured to store multiple object identifiers to cover the transport of multiple objects simultaneously.

[0037] According to an advantageous embodiment of the invention:

[0038] - the container is configured to categorize its interlocutor into a typical conversation profile and transmit the typical conversation profile to the intelligent conversational agent;

[0039] - the intelligent conversational agent being configured to adapt the semantic analysis according to the typical conversation profile.

[0040] Thanks to these technical characteristics, the intelligent conversational agent is able to target its responses according to the category of its interlocutor.

[0041] According to a particularly advantageous embodiment of the invention, the intelligent conversational agent is configured so as to limit the communication of information during the conversation to information relating to the secure transport of the object, if the typical conversation profile does not have the status of recipient of the container.

[0042] Thanks to these technical characteristics, the intelligent conversational agent can limit the dissemination of information on the object delivered to its recipient and thus prevent an actor in the delivery chain from becoming aware of its content.

[0043] According to one embodiment, the invention also provides an object transport system comprising a fleet of containers.

[0044] Thanks to these technical features, the system can benefit from the feedback of information from all the containers forming the fleet. This allows, for example, to improve responses relating to the use of reusable smart parcels for transporting objects.

[0045] According to one embodiment, the mobile container is a reusable smart packaging or package for shipping objects.

[0046] According to a preferred embodiment of the invention, the first data and the second data are so-called communication data which form a conversational exchange, and in which the first data comes from a user of the container; the processing of the second data by the container consists of communicating the second data to said user, the conversational intelligent agent is further configured so as to:

[0047] - retain contextual information from the conversational exchange,

[0048] - use contextual information during said semantic analysis. Thanks to these characteristics, a user can query the container in natural language and obtain in return help and / or details provided in natural language. In addition, the user can converse with the container through a series of chained questions / answers.

[0049] In addition, the intelligent conversational agent has the history of questions / answers exchanged with the user and can therefore refine its analysis in order to increase the accuracy of its responses.

[0050] According to an advantageous embodiment of the invention, the first data is so-called information data, of the type, an identifier, or a physical quantity, or an operating configuration.

[0051] Thanks to these characteristics, the intelligent conversational agent receives information about the context of the container. This information can be simple, coupled or composed to provide a variable level of precision on the information collected. The first data, called information, can for example include the internal temperature alone and / or external, the internal and / or external atmospheric pressure. This can also be information related to its movement, such as acceleration, or a position by satellite location. Thus, from the movement information, the intelligent conversational agent can deduce the subjection of the container to a shock, vibrations, or a fall. In the case of a fall, the height of the fall and / or the acceleration can complete the information.

[0052] Advantageously, the identifiers are identifiers specific to the application, or identifiers of a known type, such as that of a SIM ® card (an acronym for "Subscriber Identify Module"), or an NFC ® code (an acronym for "Near-Field Communication"). According to a particularly advantageous embodiment of the invention, the second data item, resulting from the semantic analysis of the first data item, called information, is a command applied by said container. Thanks to these characteristics, optionally depending on the result of the semantic analysis, a second data item is produced, then sent to the container which can then use it.Thus, the conversational intelligent agent can choose between sending a text or voice message and commanding different actions, including controlling an LED (acronym for Light Emitting Diode also known by the acronym "LED"), a siren or a speaker, a motor, a display screen, a locking / unlocking device. Thus, the conversational intelligent agent can, for example, command a flashing mode of the LEDs to attract the user's attention.

[0053] According to a preferred embodiment of the invention, the intelligent conversational agent is configured so as to:

[0054] - record the first so-called information data in a continuous learning parameter base; and

[0055] - train a model of the conversational intelligent agent with the continuous learning parameters stored in said database.

[0056] Thanks to these characteristics, the intelligent conversational agent increases its knowledge base, so as to enrich its responses and / or improve their relevance.

[0057] According to a particular aspect of the invention, the drive is automatic.

[0058] According to a particularly advantageous embodiment of the invention, the training is deep learning, in which each continuous learning parameter is associated with a weighting of positive or negative assessment of said parameter. Thus, thanks to the help of “machine learning” or “deep learning” type algorithms, the value of the learning data is qualified, improving the reliability of subsequent semantic analyses.

[0059] According to one embodiment, the container comprises a means for entering said first so-called communication data, which is a voice capture means and / or a text input means.

[0060] Thanks to these characteristics, the user can query the container directly in natural language orally and / or use an input means, such as a touch or mechanical keyboard, or a stylus associated with a handwriting recognition means, to communicate with the container. According to a particular embodiment of the invention, the input means is an independent, proximity remote control, such as a smartphone. The smartphone then communicates with a proximity wireless communication means such as a Bluetooth ®, BLE ® (English acronym for “Bluetooth Low Energy”) or Wi-Fi ® protocol.

[0061] Thanks to these characteristics, it is possible to produce containers at reduced prices and thus expand the target customer base of the product.

[0062] According to an advantageous embodiment of the invention, the system comprises a means of text transcription of said voice capture, so that the intelligent conversational agent receives a first data item in text format.

[0063] Thanks to these characteristics, using speech, the user can communicate with the container whose captured voice message is converted into a text message received by the conversational intelligent agent, and directly interpretable by a conversational artificial intelligence. According to one embodiment, the container comprises a means of sound reproduction and / or a means of displaying said second so-called communication data.

[0064] Thanks to these characteristics, the user receives a response to the query that he / she submitted indirectly to the conversational intelligent agent through the container.

[0065] According to an advantageous embodiment of the invention, the system comprises a means for voice transcription of the second so-called communication data produced in text format by the intelligent conversational agent. Thanks to these characteristics, the user receives a voice response.

[0066] According to one embodiment, the container is selected from a group comprising luggage and a reusable box.

[0067] Thus, in a new and unexpected way, a reusable box, a reusable logistics delivery packaging, an electronic logistics label intended to equip a single-use packaging and also luggage, such as a suitcase, allow a user to converse to obtain different types of information, including information facilitating the use of the container.

[0068] In a particularly advantageous embodiment of the invention, the intelligent conversational agent is integrated into a computer unit embedded in said container and / or on a central computer platform configured to communicate with said container.

[0069] Thanks to these characteristics, it is possible to have either an autonomous container with localized intelligence, or a container supported by intelligence located on a remote central computing platform. It is also possible to distribute, or even combine intelligence on both sides between the central computing platform and the container, to increase the relevance and / or speed of responses, or even distribute the processing load according to the sizing of the processing power embedded in the container.

[0070] In addition, intelligence located in each container also offers the advantage, particularly in the case of luggage, of improving confidentiality and / or respect for privacy.

[0071] Furthermore, such localized intelligence makes up for the lack of connectivity with the central computing platform when long-distance communications fail. This also helps reduce the power consumption of the long-distance communication module. This also helps manage and therefore minimize the volume of 4G data consumption.

[0072] According to a particular embodiment of the invention, the intelligent conversational agent is transferred to an externalized IT platform configured to communicate with said central IT platform.

[0073] Additionally, the conversational intelligent agent and the container are configured to communicate with each other via the central computing platform.

[0074] Thanks to these characteristics, the intelligent conversational agent is transferred to an outsourced platform, for example from a specialized supplier.

[0075] According to a particular embodiment of the invention, the transport system further comprises a means of communication with a global computer network. According to alternative embodiments, the means of communication connects the central computer platform and / or the outsourced computer platform to the global computer network. Thanks to these characteristics, the conversational intelligent agent is able to complete its learning by consulting the internet. In addition, it can enrich its responses with information collected on the global computer network, such as, for example, weather information related to a temperature peak recorded by the container in a specific location.

[0076] According to an advantageous embodiment of the invention:

[0077] - the container is configured to identify the user, categorize the identified user into a conversation profile type and transmit the conversation profile type to the conversational intelligent agent;

[0078] - the intelligent conversational agent being configured to adapt the semantic analysis according to the typical conversation profile.

[0079] Thanks to these characteristics, the conversational intelligent agent is able to provide targeted information based on the typical conversation profile and therefore the role of the user, who may have the role of a customer receiving a package, that of a logistician responsible for its delivery, or even the role of an agent in charge of the maintenance of the reusable package. Thus, the identified maintenance agent can, using voice command keywords, obtain diagnostic information, or even information or recommendations from the after-sales service manual, which are not available for other roles.

[0080] According to one embodiment of the invention, the identification of the user is completed by authentication.

[0081] Thus, thanks to this technique, it is possible to ensure, for example, that only the recipient customer can open the container upon delivery.

[0082] According to a particularly advantageous embodiment of the invention, the intelligent conversational agent is configured to choose a conversation language based on a position of the container, received in the first so-called information data. The position, which is a physical location quantity, is typically the last GPS position of the container. But this can also be deduced from other information, such as the current cellular telephone cell, or the known position of the current storage warehouse and / or its Wi-Fi® network to which the container can be connected.

[0083] Thus, the intelligent conversational agent communicates in the local language when it is available, or switches to an international language, typically English, or even Spanish or French.

[0084] Advantageously, the intelligent conversational agent uses the location position to determine local cultural criteria and adapt the responses produced to these criteria.

[0085] Certain aspects of the invention are based on the idea of ​​implementing advances in artificial intelligence in the conversational field, in particular with large language models or LLM (English acronym for "Large Language Model"), of the type of "pre-trained generative transformers", such as services like GPT3 ® and ChatGPT ® from OpenAI ®, Bert ® and Bard ® from Google ®, or even with generative artificial intelligence.

[0086] These aspects of the invention include the idea of ​​being able to quickly and easily adapt, or even update, the responses to be given to the interlocutor, depending, in particular, on the object being transported. Other aspects of the invention include the idea of ​​significantly optimizing the relevance of these responses, so as to provide a high level of precision in them.

[0087] One objective of the invention is to provide a technique that is reliable, and therefore consequently which notably reduces the risk of hallucination phenomena in the responses provided. As a reminder, in artificial intelligence, hallucination is the fact of providing a statistically plausible but erroneous response to a requester who could misleadingly consider it as a real or certain fact.

[0088] Advantageously, the invention uses augmented retrieval generation or RAG (acronym for "Retrieval Augmented Generators"). Such optimization of generative artificial intelligence is implemented using technological building blocks well known to those skilled in the art such as AutoGen ® from Microsoft ®, Langchain ® from Langchain Inc. ®, Llamalndex ® from Llamalndex ®, cited as illustrative and non-limiting examples.

[0089] Thus, the use of such techniques presents numerous advantages. The use improves the efficiency in the use of LLMs with the processing on significantly, or even much smaller, databases. This use also provides great flexibility for the creation of new corpora or the addition of new documents. It also increases the reliability of the results by better control of the knowledge base, by mastering the sources of documents used, which reduces the risk of hallucination phenomenon on the answers provided. Such use further contributes to improving confidence in the answers provided, due to the advantages previously mentioned. . Presentation of figures

[0090] Other characteristics and advantages of the invention will appear more clearly on reading the following description of embodiments of the invention, given as simple illustrative and non-limiting examples, and the appended drawings among which:

[0091] [Figure 1] Figure 1 is a schematic representation of an exemplary embodiment of an object transport system featuring a communicating container according to the invention;

[0092] [Figure 2] Figure 2 is a schematic representation of a second exemplary embodiment of smart luggage according to the invention. Detailed description of the invention

[0093] 1. Example of an embodiment of the invention

[0094] Figure 1 illustrates a first embodiment of an object transport system according to the invention, comprising a container 1, an IT infrastructure 14 and an externalized IT platform 19.

[0095] The container 1 allowing a conversational exchange integrates electronic means including an on-board computer unit 2, data storage means 3, a power supply battery 4, a microphone 5, a loudspeaker 6, a digital screen 7, one or more control buttons 8, a first communication means 9, one or more actuators not shown, the assembly assembled according to techniques known per se.

[0096] It also integrates a multitude of sensors (not shown) to obtain various information related to its storage, transport, or environment. Thus, it can obtain a GPS (Global Positioning System) position allowing its movement to be traced, movement information and the receipt of a shock, its exposure to vibrations, or even recording a fall. The container 1 is also capable of recording internal and / or external information such as temperature and atmospheric pressure. This various information also includes an identifier specific to the container 1, an identifier specific to the computer unit 2, a SIM (Subscriber Identify Module) card identifier, a near-field communication code known by the acronym N FC (Near-Field Communication), or any other identifier known per se to those skilled in the art.The IT infrastructure 14 includes a computer platform 15, a database 16 and a second communication means 17 for communicating via the first communication means 9, with the computer unit 2 of the container 1.

[0097] Furthermore, the IT infrastructure 14 is capable of communicating using a third means of communication 20 with the outsourced IT platform 19.

[0098] In this embodiment, the outsourced IT platform 19 includes a conversational artificial intelligence module, typically software, which constitutes a conversational agent 18 configured to: a) receive captured data, transmitted by the container 1 via the IT infrastructure 14; b) perform a semantic analysis of the captured data; c) produce data resulting from the semantic analysis; and d) transmit via the IT infrastructure 14, the resulting data to the container 1;

[0099] A conversation begins when the captured data is a first so-called communication data, and the resulting data is a second so-called communication data. In this case, the first communication data and the second communication data are stored by the conversational agent 18 according to methods known per se, to retain contextual information. This is then a history of the conversation composed of this contextual information which is enriched over the course of the exchanges.

[0100] To initiate a conversation, a user presses the button 8 of the container 1. The computer unit then triggers an audio capture using the microphone 5 and records the voice message emitted by the user. Then, it transmits this voice message, called the first voice message captured, to the computer infrastructure 14. The latter, equipped with voice-to-text conversion means known by the English terminology of "speech to text" and "text to speech", is responsible for transcribing the first voice message into a first corresponding text message. The computer infrastructure 14 then transmits this first text message to the conversational agent 18 on the outsourced computer platform 19.

[0101] In a conversation, the resulting data, which is a second text message, is first received by the IT infrastructure 14. This converts the second text message into a second voice message and transmits it to the container 1. Then, the container 1 broadcasts the second voice message through the loudspeaker 6, to the user, who can continue the conversation. Thus, by the sequence of these exchanges, the container 1 supports the conversation with the user.

[0102] Furthermore, being equipped with other sources of information specified above, the container 1 is also configured to transmit at least one of the various information available as a captured data message. For example, the captured data transmitted carries the information of its GPS position. Or the captured data contains a combination of information, such as the identifier, the internal temperature and the external temperature measured by the container 1.

[0103] When the captured data is therefore not the user's voice message but so-called information data, the resulting data is a command transmitted to the container 1 which applies it to at least one of the actuators. For example, the resulting command commands the unlocking of the container 1, authorizing its opening by the user. However, the user's voice message can be interpreted by the conversational agent 18 as a command to control one of the actuators. However, depending on the configuration of the container 1 and / or the conversational agent 18, the user cannot necessarily act on all the actuators equipping the container 1.

[0104] In the context of the object transport system, using a fleet of containers 1 , it is all the data sent by the fleet of containers to the conversational agent 18 which contributes to the enrichment of the model.

[0105] Furthermore, contextualizing container 1 allows the conversation to be oriented in order to provide the user with the most appropriate information possible in relation to container 1 and in particular its possibilities.

[0106] In addition, contextual information is subjected to the learning of the model of the conversational agent 18, so as to instill in it a corpus of initial conditions. Typically, this contextual information relates to the container 1, its description, its instructions for use, possible uses, frequently asked questions submitted by users, the commercial offer for transporting objects using the container 1, or information on the company managing the fleet of containers 1 forming the object transport system.

[0107] To do this, the conversational agent 18 is subjected to a specialization of its model in which it is provided with a series of key words and the associated semantics to form the corpus of initial conditions.

[0108] As an example, learning the events associated with a shipment can be done using the following illustrative pseudo-code:

[0109] “Each stored event is composed of type, value where date is the date of the event, type is the type of the event and value is defined per type as follows:

[0110] - 'sender': name, address

[0111] - 'recipient'mame, address - 'startshipment': date, latitude, longitude

[0112] - 'endShipment': date, latitude, longitude

[0113] - 'sensors':date, 'externalTemperature', measure (in °C), 'internalTemperature', measure (in °C), 'externalHumidity', measure (in%), 'internalHumidity', measure (in%), 'pressure', measure (in hPa)

[0114] - 'geolocation': date, latitude, longitude

[0115] - 'warehouse': enterDate, exitDate, latitude, longitude, name

[0116] - 'flight': fromDate, fromLatitude, fromLongitude, fromAirport, toDate, toLatitude, toLongitude, toAirport

[0117] - 'alert': date, latitude, longitude, subtype (in 'externalTemperature', 'internalTemperature', 'externalHumidity', 'internalHumidity', 'freeFall'), measure

[0118] -'events': You will return every events containing the keyword 'store' as a list without displaying the keyword itself. When 'event' is followed by a type, return every events of the defined type as a list instead of all 'store' keywords and return the closest city instead of raw location coordinates. »

[0119] L’apprentissage de la sémantique associée peut être réalisée de la façon suivante :

[0120] « A duration between two events event A and event B is defined as the difference between the date of the event B and the date of the event A.

[0121] The duration of an event is defined as the difference betwen the exitDate and the enterDate of the event or the toDate and the fromDate of the event»

[0122] Le contenant 1 est par exemple un colis ou un emballage d’expédition conforme à l’invention.

[0123] A principle of contextualization centered on the use of the container has previously been described. In this example, the corpus of initial conditions contains all or at least information about the container deemed useful to answer the user's questions on this subject. In this case, the contextual information then includes this type of corpus of initial conditions, specialized on information relating to the container and the history of questions / answers exchanged with the interlocutor or user. Furthermore, a corpus of initial conditions has also been described which concerns the shipment of the container, notably taking into account the information sent by container 1 and / or by the fleet of container 1.

[0124] However, this principle of contextualization is only an illustrative example that should not be understood as a definitive assertion. Thus, other types of corpus are included within the scope of the invention.

[0125] According to a particularly advantageous embodiment, the container 1, in the form of an intelligent reusable shipping package, has information on the object or product that it contains and transports. Thus, the contextual information can relate to any information dedicated to the object or product to be transported, contributing to forming a specific corpus of initial conditions, used by the model of the conversational agent 18.

[0126] By way of illustrative and non-limiting examples, this information on the object shipped may include the detailed characteristics of the object, whether its technical or aesthetic characteristics, its advantages, possibly its disadvantages or limitations, but also its implementation or detailed instructions for use of the object.

[0127] In other words, the IT infrastructure 14 and / or the container 1 stores an object identifier for each of the objects, present during the shipment, in the receiving volume of the container 1. By communicating each object identifier to the model of the conversational agent 18, the latter determines in a library, also called a set of initial condition corpora, the corpus of contextual data to be used. It is understood that for the example of this embodiment, a corpus of initial conditions is dedicated to a particular object. In this case, the corpus of initial conditions is then specialized. In other words, the conversational agent 18 chooses from a set of initial condition corpora, or library which covers a set of objects, without necessary link between them the corpus of initial conditions in relation to the transported object using the object identifier serving as a selection criterion.

[0128] It is important to understand that a corpus of initial conditions can concern a broader scope than a single object, for example including the description of a range of products or a product catalog bringing together all the products available for sale from a seller. It is clear that the information used to enrich a corpus can also come from the instructions for use of a product already mentioned, from a list of recommended products or services directly or indirectly linked to the product transported, such as the recommended maintenance conditions or maintenance operations.

[0129] Thus, thanks to the knowledge of the product transported, via the object identifier on the one hand and the contextual information specific to this product aggregated in the model of the conversational agent 18, on the other hand, the container 1 can support a conversation on the object that it transports. It can inform a recipient about the product that it receives, on a knowledge base much superior to that assimilated by a human interlocutor, in the role of a commercial agent, a technical interlocutor or an after-sales service interlocutor for example. In the case where the product to be delivered is a gastronomic product, such as a grand cru wine, the customer recipient, upon receipt during a conversational exchange, can obtain advice and information on the wine received, such as its organoleptic properties, or any other characteristics, the pairings with food or even the storage period, the storage conditions and consumption of this grand cru.Thus, this information or conversational exchanges can be advice on use, arguments or advantages, or even congratulations on the relevant choice of purchasing a product.

[0130] To do this, for example, the container 1 receives the identification information of the product it contains from a remote server, typically the IT infrastructure 14. Alternatively, using its embedded image sensor directed towards the interior volume of the container capable of receiving the product or object, the container captures at least one image so as to identify the product or object. This identification is obtained by any means known to those skilled in the art, in particular image recognition techniques carried out by the IT infrastructure 14 for example. Any other known method is applicable to allow the container to recover the identifier of the object it transports, in particular via a label associated with the object which emits its identifier by a near-field communication technique.

[0131] Thus, the container 1 and / or the IT infrastructure 14 are able to correlate the contextual information corresponding to the product present in the container 1. This identifier is coupled by the container 1 and / or the IT infrastructure 14 to the first so-called communication data received by the conversational agent 18. In summary, the question asked by the interlocutor of the container 1 is transmitted in the form of the first data associated with the object identifier to the conversational agent 18 which provides a response via the second so-called communication data. This response provides information on the object, product, transported when the question relates to this subject.

[0132] It is perfectly clear and understood within the framework of this invention that at any time during the journey, that is to say the transport of the container 1, the information sent back by the container and aggregated in the model of the conversational agent 18 can be, in whole or in part, communicated to the participants in the logistics chain, to the sender or to the recipient. However, in the event of necessary confidentiality, the information remains shared between the current user and the conversational agent 18. The trace of this contextual information is then not kept at the end of the delivery. In other words, the identifier of the object in this particular circumstance is erased in the place or places where it was stored. The link between this object identifier and the customer during this delivery does not persist and therefore no longer exists.

[0133] Furthermore, only information relating to the security of goods and people is available for a conversation with the various actors in the supply chain during transport. Apart from this security information, all other information is only available to fuel a conversation with the clearly identified recipient of container 1, a reusable smart package. In other words, the type of interlocutor who starts a conversation is identified, then classified or categorized in a typical conversation profile, to ensure that he has the status of recipient of container 1, before communicating to him information revealing the nature of the transported object. In other words, only the recipients of the package obtain information about the object they receive during a conversation.The recipients are therefore the only ones, once the package has been shipped, to be able to obtain all the information available in the corpus of initial conditions corresponding to the product delivered.

[0134] In a variant of the previous embodiment, the conversational agent 18 chooses several corpora of initial conditions in the library, forming a group of selected corpora. This choice is made according, here too, to a selection criterion which is the received object identifier. This choice criterion can also be a function of the question asked, or according to a combination of the question and the object identifier coupled with the question. It is clear that when the question asked is taken into account in the selection criterion, it is the semantic analysis of the latter which enters into the parameters of the selection criterion. Furthermore, the choice of several corpora of initial conditions is followed by a step of selecting a part of the data in the group of corpora which forms a subset of extracted data.In other words, this step corresponds to a filtering of the data present in the corpus group, using a filter based on the previously defined selection criterion. In this variant, it is ultimately this subset of extracted data correlated with the history of questions / answers exchanged with the user that the conversational intelligent agent uses to produce its response.

[0135] Consequently, the use of a library or set of corpora makes it possible to adapt the implementation strategy and thus choose between pooling or partitioning the data from the different corpora depending on the confidentiality requirements required for the products transported.

[0136] 2. Other examples of embodiments of the invention

[0137] In Figure 2, the container 1 is represented as a piece of luggage, here a suitcase 10 forms an object transport system according to a second embodiment of the invention.

[0138] The container 1 also comprises electronic means including the on-board computer unit 2, the data storage means 3, the power supply battery 4, the microphone 5, the loudspeaker 6, the button 8, assembled according to techniques known per se. The on-board computer unit 2 of the suitcase 10 is configured to capture, using the microphone 5, requests formulated by the user of the suitcase 10 and to emit responses via the loudspeaker 6. Thus, the microphone 5 and the loudspeaker 6 form an interface with the user of the suitcase 10.

[0139] Furthermore, the on-board computer unit 2 is configured to execute the conversational intelligent agent, and thus autonomously maintain a conversation with the user of the suitcase 10. Furthermore, the on-board computer unit 2, thanks to the communication means 9, is able to receive a specialized corpus chosen according to the intended use of the suitcase 10. For example, this corpus is specialized according to a particular journey, with information on the scheduled transit locations such as stations, airports, transport infrastructure of a city with information on the map of the metro stations, etc. Thus, the user can, thanks to a conversation with the suitcase 10, obtain information to find his way through the maze of corridors of these transit locations.

[0140] In a variant of the second embodiment, the on-board computer unit 2 communicates via the communication means 9 with the intelligent conversational agent 18 remoted on the computer infrastructure 14, or on the externalized computer platform 19, not shown.

[0141] Within the framework of the invention, it is obviously intended to apply the principle, described in the previous chapter, of conversation about the object contained in a container 1 of the intelligent shipping package type, to the luggage illustrated by figure 2.

[0142] Thus, based on the same operation, the owner of the luggage has the role of recipient of a package and becomes the interlocutor of the intelligent conversational agent 18.

[0143] In this case, the owner can provide information on the baggage about the object or product it contains when it cannot be obtained using the methods previously described. Thus, the object identifier is known, usable and can be coupled with the first communication data as already described.

[0144] When the luggage is used for the presentation by its owner, a seller of products, such as new samples of medical prostheses or new medications, the intelligent conversational agent 18 can, by choosing the appropriate set of initial conditions, support a conversation with a health professional, doctor, surgeon, to communicate to him all the detailed information on the product presented. There are then no more questions left unanswered that the representative could not have answered immediately. This time saving increases the relevance of sales interventions, and avoids having to again call on overworked professionals to provide them with the details that the seller could not have provided himself.

[0145] 3. Other optional features and advantages of the invention

[0146] In variants of the embodiments of the invention detailed above, it may also be provided: to record, at the level of the conversational agent 18, the captured data in a database for storing continuous learning parameters, when the captured data is at least one of the various pieces of information, so as to train a model of said conversational agent 18, according to a technique known in the English terminology of “machine learning”. Thus, the conversational agent 18 is able to train its model on the one hand and it is configured to train it automatically on the other hand;to associate the captured data with a weighting of positive or negative assessment, taken into account by the conversational agent model 18. Thus, the training of the model is a deep learning technique, known under the English terminology of “deep learning”. This increases the statistical analysis performance of the model and consequently the relevance of conversational exchanges with the user; to integrate into the container means of identification and / or authentication of the user to determine his role and consequently the conversation profile to adopt. Identification can be obtained by any known method, such as a badge, a means of voice recognition, etc.;to integrate the voice-to-text conversion means into the container 1, or to transfer these means to the outsourced platform, or to a specialized supplier. to integrate the conversational agent 18 into the IT infrastructure 14, in particular optimizing response times; to embed the conversational agent 18 in the IT unit 2 of the container 1 serving as shipping packaging, and thus give it autonomy to converse with the user; for the conversational agent 18 embedded in the IT unit 2, to receive possibly weighted continuous learning parameters, emitted by the IT infrastructure 14, the learning parameters being based on feedback from a fleet of containers 1 in operation; to specialize the intelligent conversational agent so that it adapts its semantic analysis according to the current location of the container;to transmit an instruction, in other words a voice message to the user, following the analysis of captured data, for example to issue a message requiring more precautions in handling the container 1 when a fall is detected; to use a digital touch screen as an input and / or output interface between the container 1 and the user; to integrate the first means of communication 9 with the electronic means of the suitcase 10; to create dedicated fleets per customer of a logistics carrier, with a partitioning of the contextualization and therefore of the initial corpus used by the conversational agent 18 according to the specificities of the customer; to use, for example, a serial number of the transported object as an identifier of the object;to provide a specialized corpus based on a particular journey, with information on scheduled transit locations such as stations, airports, and a city's transport infrastructure, to facilitate the transport of a package 1; that the object identifier is considered as a second identifier within the meaning of the invention, when all the other identifiers previously cited are considered as the first identifier, such as, for example, the identifier of a SIM ® card (an acronym for "Subscriber Identify Module"), or that of an NFC ® code (an acronym for "Near-Field Communication"); that all the identifiers previously defined and / or cited, in particular the identifiers specific to the application, are usable by the intelligent conversational agent; that a library or set of initial condition corpora is not necessarily fixed, but can dynamically be constituted and / or enriched according to needs;that each user of the package, throughout the logistics chain, from the sender to the recipient, is an interlocutor wishing to maintain a conversation with the intelligent conversational agent.;

[0147] The technique described above for creating an object transport system enabling conversational exchange can be used with different types of receptacles, for example to create a container of the goods transport container type.

[0148] It should be understood that the elements shown in the figures may be implemented in various forms of hardware, software, or combinations thereof. Preferably, these elements are implemented in a combination of hardware and software on one or more suitably programmed general-purpose devices, which may include a processor, memory, and input / output interfaces.

[0149] The present description illustrates the principles of the present disclosure. It will therefore be appreciated that persons skilled in the art will be able to devise various arrangements which, although not explicitly described or shown herein, embody the principles of the disclosure and are included within its scope. Thus, according to the principle of the invention it should be understood that the library(s) or set of initial condition corpora, make it possible to select at least one corpus of contextual data which serves as support for the conversational agent to support a conversation with the user. This or these libraries can be enriched and / or specialized according to the needs of the operator of the object transport system, or other actors such as the sender and thus broaden the possible discussion contexts beyond the examples previously described.

[0150] All examples and conditional language cited herein are intended for educational purposes to assist the reader in understanding the principles of the disclosure and the concepts contributed by the inventor to advance the art, and should be construed as not being limited to these specifically cited examples and conditions.

[0151] Furthermore, all statements of principles, aspects and embodiments of the disclosure, as well as specific examples thereof, are intended to encompass their structural and functional equivalents. Furthermore, it is intended that such equivalents include both currently known equivalents and equivalents developed in the future, i.e., any developed element that performs the same function, regardless of its structure.

[0152] Thus, for example, those skilled in the art will understand that the block diagrams presented herein represent conceptual views of illustrative circuits implementing the principles of the disclosure. Likewise, it will be appreciated that all flowcharts, flow diagrams, and the like represent various processes that may be substantially represented on a computer-readable medium and thereby executed by a computer or processor, whether or not such computer or processor is explicitly shown. To the extent that embodiments of the invention have been described as being implemented, at least in part, by a software-controlled data processing apparatus, it will be understood that a non-transitory machine-readable medium carrying such software, such as an optical disk, magnetic disk, semiconductor memory, or the like, is also considered to represent an embodiment of the present invention.

[0153] The functions of the various elements illustrated in the figures may be performed by the use of dedicated hardware as well as hardware capable of executing software in association with appropriate software. When performed by a processor, the functions may be performed by a single dedicated processor, by a single shared processor, or by a plurality of individual processors, some of which may be shared. In addition, the explicit use of the term "processor" or "controller" should not be construed as referring exclusively to hardware capable of executing software, and may implicitly include, without limitation, a digital signal processing (DSP) processor, read-only memory (ROM) for storing software, random access memory (RAM), and a nonvolatile storage medium.Additionally, some elements may be independent or grouped together, for example in a microcontroller (MCU), or a system-on-chip (SoC).

[0154] Other hardware, conventional and / or custom, may also be included. Similarly, any switches shown in the figures are conceptual only. Their function may be performed by the operation of program logic, by dedicated logic, by the interaction of a program control and dedicated logic, or even manually, the particular technique being selectable by the performer as more specifically understood in the context. In the claims herein, anything expressed as a means of performing a specified function is intended to encompass any manner of performing that function, including, for example, a) a combination of circuit elements that performs that function or b) software in any form, including, therefore, firmware, microcode, or the like, combined with suitable circuitry for executing that software to perform the function.The disclosure as defined by such claims is that the functionalities provided by the various means recited are combined and brought together in the manner provided by the claims. It is therefore considered that all means which can provide these functionalities are equivalent to those presented here.

Claims

CLAIMS 1. Object transport system enabling a conversational exchange with an interlocutor, comprising: - at least one intelligent mobile container for transporting objects capable of containing an object and capable of conversing with an interlocutor, the package being configured so as to capture a first piece of data; - an object identifier of the object contained by said container (1) during the transport of said object; and - a conversational intelligent agent (18) configured to: a) receive the first data item coupled with said object identifier contained during transport in the container, the first data item being transmitted by said container, the object identifier being further transmitted so as to enable a conversational exchange relating to the transported object; b) perform a semantic analysis of said first data item received, contextualized in light of the object identifier coupled with said first data item, forming a semantic result; c) produce, optionally, a second contextualized data item, resulting from the semantic analysis; d) transmit, optionally, said second data item to said at least one container in response to said first data item; said container being further configured so as to process the second data item received.

2. Object transport system according to claim 1, wherein the conversational intelligent agent (18) is configured to: - select a first corpus of initial conditions based on a selection criterion, from a library or set of corpora of initial conditions; - use the first corpus of initial conditions when producing the second contextualized data.

3. Object transport system according to claim 2, in which the selection criterion is a function of the object identifier, and / or the semantic result of the analysis of the first data.

4. Object transport system according to claims 2 or 3, wherein said first corpus of initial conditions comprises information relating to said transported object, to enable the conversational intelligent agent to maintain a conversation relating to said object.

5. Object transport system according to one of claims 1 to 4, in which the conversational intelligent agent (18) is configured to select at least one second corpus of specialized initial conditions according to the selection criterion, in the library, the first corpus and the at least one second corpus forming a group of corpora.

6. Object transport system according to claim 5, wherein the conversational intelligent agent is configured to: - extract a subset of data from the corpus group using a filter based on the selection criterion; - use said subset of data in the production of the second contextualized data.

7. Object transport system according to one of the preceding claims, in which the container (1) is configured to: - store the object identifier of the object present in the container, and - attach the stored object identifier, coupled with the first data transmitted to the conversational intelligent agent (18).

8. Object transport system according to one of the preceding claims, comprising a fleet of containers.

9. Object transport system according to one of the preceding claims, in which: - the container is configured to categorize its interlocutor into a typical conversation profile and transmit the typical conversation profile to the intelligent conversational agent; - the intelligent conversational agent being configured to adapt the semantic analysis according to the typical conversation profile.

10. Object transport system according to the preceding claim, in which the conversational intelligent agent (18) is configured so as to limit the communication of information during the conversation to information relating to the secure transport of the object, if the typical conversation profile does not have the status of recipient of the container (1).

11. Object transport system according to one of the preceding claims, in which the first data and the second data are so-called communication data which form a conversational exchange, and in which said first data comes from the interlocutor of said container; said processing of the second data by the container consists of communicating said second data to said interlocutor, the conversational intelligent agent is further configured so as to: - retain contextual information from the conversational exchange, - use contextual information during said semantic analysis.

12. Object transport system according to one of the preceding claims, in which the intelligent conversational agent is configured to choose a conversation language based on a geographical position of the container.

13. Object transport system according to one of the preceding claims, in which the conversational intelligent agent is integrated into an on-board computer unit (2) of said container and / or on a central computer platform (14) configured to communicate with said container.

14. Object transport system according to claim 13, in which the intelligent conversational agent is transferred to an externalized computing platform (19) configured to communicate with said central computing platform, the intelligent conversational agent (18) and the container (1) being configured to communicate together via the central computing platform.

15. Object transport system according to one of the preceding claims, in which the mobile container is a reusable intelligent package (1) for shipping objects.

Images