System and process for optimizing the Quality of Experience (QoE) of the user on the go

Abstract

The invention relates to a system (100) and method (100') for optimizing the Quality of Experience (QoE) of a mobile user ("QoE-d"), comprising a proactive adaptive system (300) for managing connection data flow based on an updated prediction of the Quality of Service established by a predictive system (200), coupled with recommendation, reservation, and connection access coordination systems. This system (100) and optimization method (100') is particularly well-suited to communication applications used in mobile environments. By leveraging the analysis of network distribution and availability data, the system optimizes communication stability and offers an innovative solution to the limitations encountered in mobile environments.

Description

Title of the invention: System and method for optimizing the Quality of Experience (QoE) of the user on the go. TECHNICAL FIELD OF THE INVENTION

[0001] The present invention relates to the field of telecommunications, and more particularly, a computer-implemented system for optimizing the Quality of Experience (QoE) of the user on the move. STATE OF THE ART

[0002] Nowadays, personal and professional communications play a central role in everyone's daily life, making it essential to maintain a stable connection under all circumstances. However, the rise of teleworking and the increase in business travel imply, among other things, infrastructural constraints (such as physical obstacles), irregular network distribution, and instances of network congestion, which can impair the Quality of Service (QoS) observed along a given route. Under such conditions, users frequently experience instability in their communications and therefore a degradation of the Quality of Experience (QoE).

[0003] This is why it becomes crucial to establish a Quality of Service (QoS(n)) prediction for a planned route in order to optimize the QoE for the user in motion. The prior art offers a predictive system for the performance of communication means (network providers) in order to select the most efficient communication means, while the main objective of the present invention is to adapt the connection data flow based on the QoS(n). OBJECT AND SUMMARY OF THE INVENTION

[0004] The present invention proposes a system for optimizing (100) the Quality of Experience (QoE) of the user in motion, comprising a predictive system (200) of the QoS in mobile condition, for a future route, from which an adaptive system (300) adapts a data stream.

[0005] To achieve this objective, said QoE optimization system (100) includes: a. A predictive system (200) of the future QoS “QoS(n)”; b. A recommendation (212') and booking (214') system for the best connection slot; established from the QoS(n); c. A proactive adaptive (300) data flow system, based on the predicted QoS (QoS(n)); d. A transition stabilization module (310) allowing the integration of an observation delay (310') preceding the adaptation of the connection data flow; e. A centralized management module (222) allowing coordination of access to the connection (222') from the QoS(n); f. A telecommunications system integrating all the systems described, as well as a messaging system between users. Description of the invention

[0006] The invention relates to a system for optimizing (100) the Quality of Experience (QoE) of the user on the move ("QoE-d"), comprising an adaptive data flow system (300) based on the Quality of Service ("QoS(n)") prediction established by a computer-implemented predictive system (200). Said QoE-d optimization system (100) is implemented according to a QoE-d optimization method (100') comprising: a. A QoS prediction phase (200') in mobile conditions, for a future route, established by the predictive system (200); b. An adaptation phase (300') of the connection data flow from the QoS prediction established by the adaptive system (300).

[0007] Said optimization system (100) attempts to compensate for variations in connection conditions observed during movements. Brief description of the drawings

[0008] Fig. 1 illustrates one embodiment of the system (100) and method (100') for optimizing QoE-d;

[0009] Fig. 2 illustrates an embodiment of the "prediction phase" (210') of the prediction phase (200'), aimed at producing, before the movement (at t(0)), an Initial Prediction of the QoS: PLQoS(n) (with n going from 0 to N, 0 corresponding to the start of the journey, N corresponding to the end of the journey and with a step ô);

[0010] Fig. 3 illustrates an embodiment of the "update phase" (220') of the prediction phase (200'), aimed at producing, during the movement (at t(k)), an Updated Prediction of the QoS: PA-QoS(n) to come (with n going from k to N, k corresponding to a moment of the journey from the start (so k > 0, N corresponding to the end of the journey, and with a step ô);

[0011] Fig. 4 illustrates one embodiment of the proactive adaptive connection data flow system (300).

[0012] The embodiments shown in the figures of the accompanying drawings are merely illustrative examples and do not limit the invention. The same reference numerals in the figures designate similar elements. DETAILED DESCRIPTION

[0013] Note: a. Quality of User Experience on the Move “QoE-d” b. Quality of Service (QoS) c. Predicted Quality of Service along a future route “QoS(n)” d. The Initial Prediction at t(0) of the QoS(n): “PLQoS(n)” (with n ranging from 0 to N) e. The Updated Prediction at t(k) of the QoS(n): “PA-QoS(n)” (with n ranging from k to N) f. Initial data: “x(t0)” g. Real-time data: "x(tk)"

[0014] The predictive system (200) implemented by computer predicts the QoS based on initial data x(t0) as well as real-time data x(tk).

[0015] The data collected includes, in particular: a. Data relating to network distribution (coverage), b. Data relating to network availability (saturation), c. Data relating to Quality of Service (QoS) (bandwidth quality, jitter, packet loss, latency...), d. Data relating to the Quality of User Experience (QoE), e. Geolocation data, f. Data relating to the internal state of the connected device, g. etc...

[0016] It should be noted that, depending on their nature, data are collected at varying frequencies. For example, GPS data is collected in real time, while network information is updated according to availability.

[0017] This data is collected, stored, and then processed using a predictive analytics algorithm based on machine learning or statistical modeling techniques. This predictive analytics model is specifically developed and adapted to the mobility context.

[0018] The prediction phase (200') implemented by the predictive system (200) comprises two phases: a. A forecasting phase (210') [Fig.2], preceding the movement (tO), aimed at producing, for a future journey, an initial QoS prediction "PL QoS(n)" (with n ranging from 0 to N). This forecasting phase is implemented by a "forecasting module" (210). b. An update phase (220') [Fig.3], during the movement (tk), aimed at producing, for the current journey, an updated prediction of the QoS " PA-QoS(n) » (with n ranging from k to N). This update phrase is implemented by an “update module” (220).

[0019] Said forecasting phase (210') [Fig.2] is based on the analysis of at least one of the initial data x(t0), in order to produce an initial prediction of the QoS “PLQoS(n)”.

[0020] Said forecasting phase (210') comprises the following steps: a. Collection of initial data "x(t0)": b. Retrieval of information relating to the desired route i(t0) (departure points, destination, estimated times, ...) c. Deduction of the initial data x(t0), which notably include: i. Past performance of network infrastructures; ii. Mapping of coverage and saturation areas; iii. Information provided by specialized APIs (ARCEP, etc.). d. Analysis: The initial data x(t0) are processed by a predictive model specifically designed to anticipate future network performance in the context of movement. This model relies on machine learning algorithms trained on collected, processed, and then stored data. e. Production of initial QoS predictions “PLQoS(n)”: The system generates an initial QoS prediction “PLQoS(n)” for each segment of the route. f. Storage: The PLQoS(n) and the initial data x(t0) are stored to serve as a reference during the update phase (220') and as a basis for calculating future PLs

[0021] Said update phase [Fig.3] is based on the analysis of at least one of the current data x(tk), which, combined with the PLQoS(n) as well as the initial stored data x(t0), offer an updated prediction of the future QoS “PA-QoS(n)”.

[0022] Said update phase (220') comprises the following steps: a. Collection of current data "x(tk)" which includes, in particular: i. GPS position and speed of movement; ii. Current network parameters: throughput, latency, utilization rate; iii. Unforeseen one-off events (weather, congestion, delays...); iv. Data relating to the internal state of the phone. b. Updating predictions: The system combines current data collected in real time x(tk), initial data x(t0) from the forecasting phase, and PLQoS(n). By integrating this new data, the The predictive model is re-executed to continuously update the predictions. This dynamic update, whether rolling or performed at predefined intervals "ô", allows for refining the predicted QoS for upcoming temporal and geographical segments of the route. c. Production of updated QoS predictions “PA-QoS”: The system generates an updated QoS prediction “PA-QoS(n)” at future times (with n ranging from k to N, with a step ô). d. Storage: The PA-QoS(n) and current data x(tk) are stored to serve as a basis for calculation for future predictions.

[0023] A particular embodiment of the prediction phase (200') can be illustrated by the following scenario: a. Thursday, November 10: A user plans a train trip from Paris to Marseille on Tuesday, November 15, for a business meeting with a scheduled video conference. They consult the application which, at the end of the forecast phase (210') and the prediction phase (200'), indicates a PLQoS(n) that appears to be reduced between 10:30 AM and 1:00 AM. The user can therefore schedule their video conference outside of this period of low connectivity. b. Tuesday, November 15: During the journey, the system updates in real time and detects an unforeseen event that could disrupt the journey. A PA-QoS(n) is then generated, allowing the user to move their videoconference outside the period of low coverage.

[0024] The adaptation phase (300') of the data flow is implemented by an adaptive system (300) comprising an adaptation module allowing the comparison of the updated QoS prediction to fixed or continuous thresholds (denoted "A", "B", "C", and "D"). When said level of the PA-QoS(n) crosses one of these thresholds, the system triggers a decision process aimed at adjusting the data flow to the PA-QoS(n).

[0025] Figure 4 illustrates the steps of a particular embodiment of said adaptation phase (300'): a. PA-QoS(n) greater than a first threshold "A"? i. YES: the flow of data exchanged remains at maximum: simultaneous transmission of video and audio data. ii. NO: see step 2 b. PA-QoS(n) greater than a second threshold "B"? i. YES: the flow of data exchanged is reduced: the transmission switches to a graphic avatar mode associated with audio. ii. NO: see step 3 c. PA-QoS(n) greater than a third threshold "C"? i. YES: the data exchanged is further reduced: transmission is limited to an audio stream which may be accompanied by an image ii. NO: see step 4 d. PA-QoS(n) greater than a fourth threshold "D"? i. YES: the flow of exchanged data is reduced to a minimum: speech transcribed into text for the data sent then reception of the data in text form and / or transcribed into sound form. ii. NO: Connection interruption

[0026] Unlike reactive approaches to data flow adaptation based on a real-time measured QoS level, the adaptive system (300) adopts a proactive approach based on the updated QoS prediction "PA-QoS(n)". This approach anticipates QoS(n) variations within a given timeframe, enabling the transmission and execution of the flow transition instruction before the predicted variation occurs and impacts the transmission of the transition instruction. The main advantage of this proactive approach is maintaining smooth communication and minimizing information loss, all within a general framework for optimizing the Quality of Experience for a user on the move, or "QoE-d".

[0027] In a particular embodiment, said proactive adaptive system (300) incorporates a transition stabilization module (310) allowing for an observation delay (310') between each data stream transition, in cases where the PA-QoS(n) oscillates rapidly around the thresholds (A, B, C, D). This observation delay (310') reduces unnecessary adjustments caused by temporary or transient fluctuations in the QoS(n) in order to optimize the stability and fluidity of the experience, while optimizing bandwidth consumption and limiting the loads associated with processing and executing adjustments.

[0028] In this particular embodiment, when a variation in the PA-QoS(n) would theoretically justify a transition to another data flow level, the system applies an observation delay (310') before performing said transition to ensure that the variation in the PA-QoS(n) is sustained and not transient. If, during this delay (310'), the PA-QoS(n) returns to a value corresponding to the current flow level, the system cancels or suspends the initially planned transition.

[0029] Steps in this particular embodiment of the invention: a. Detection of a variation in PA-QoS(n) (PA-QoS(n) crosses a threshold); b. Waiting for an observation period (310') of the evolution of the PA-QoS(n); c. Decision to validate or reject the transition: variation maintained during the observation period (310')? i. YES: Transition to the new flow level completed ii. NO (oscillations): transition to the new flux level cancelled.

[0030] In this particular embodiment, said observation period (310') can be configured according to: a. Predefined QoS(n) thresholds: A longer delay can be applied for transitions to reduced flow levels; b. Network environment: In environments with high QoS(n) variability, longer delays can be defined; c. User behavior: If the frequency of data flow level transitions becomes annoying for the user, the delay can be adjusted.

[0031] : The optimization process (100') of QoE-d as described in this document includes the following phases: a. QoS prediction (200') by a predictive system (200); b. Adaptation (300') of the connection data stream from the PA-QoS(n) by an adaptive system (300).

[0032] In a particular embodiment of the optimization process (100') described herein, the data processed and adjusted within the framework of the adaptive system (300) may correspond, among other things, to the data transmitted and received, within the framework of a telecommunication carried out on a web-PSTN (Real-Time Communication) communication platform, or on any other system allowing similar exchanges.

[0033] Furthermore, said data flows may, among other things, relate to: a. VoIP calls, for which the algorithm optimizes the allocated bandwidth based on network quality predictions, thus ensuring smooth communication even in degraded conditions. b. Video streaming services, where the quality of the video stream is dynamically adjusted based on predictions, offering an optimal user experience even in the event of network degradation. c. Environments requiring low latency, such as online games or drone operators, where the algorithm anticipates latency spikes and adjusts connection parameters accordingly.

[0034] In a particular embodiment, the QoE(d) optimization process (100') incorporates a centralized management module (CMM) (222) enabling coordination of connection access (222'). Connection access is thus limited to priority users in cases where the PA-QoS(n) is deemed insufficient to ensure satisfactory connection access for all users.

[0035] Said coordination of access to the connection (222') is based on PA-QoS(n).

[0036] Said coordination of access to the connection (222') is particularly suited to Environments in which access to the connection is managed centrally. These environments include, in particular, trains and airplanes.

[0037] In this particular embodiment, coordination can be carried out by a central management module (CMG) (222), such as a "Modman" (Modem Manager present in aircraft) to which the devices are connected.

[0038] In the event of a decrease in PA-QoS(n), the MGC (222) distributes access to the connection to users categorized as having a "priority" status.

[0039] In this particular embodiment, the MGC (222): a. Receives the PA-QoS(n) b. Compare the PA-QoS(n) to a threshold above which it is possible to ensure satisfactory connection access for all users: PA-QoS(n) > threshold? c. Makes a decision based on the result of the comparison: i. YES uniform distribution ii. NON-consultation of passenger status > coordinated distribution

[0040] The statuses of priority passengers are recorded in the cloud from which the MGC (222) retrieves the information.

[0041] Among the priority users, we find, for example: a. Within the framework of air transport systems: cockpit members; the cabin crew; passengers who have subscribed to a "premium" service, b. Within the framework of rail transport systems: passengers who have subscribed to a "premium" service; passengers located in "priority carriages"; etc...

[0042] In a particular embodiment, the entire optimization process (100') as described can be implemented within a telecommunications software platform (PLT), in particular compatible with WebRTC technologies.

[0043] Said PLT can integrate the entire optimization process (100') as described in this document, as well as functionalities relating to: a. Planning (reservations (214') from recommendations (212') of the best connection slots, established from the PLQoS(n)); b. Telecommunications;

[0044] The advantage of said PLT is to offer an optimized QoE-d, despite the QoS variations inherent in a displacement.

[0045] The telecommunications software platform is compatible with various "connected" devices, such as smartphones, tablets or laptops.

[0046] Said QoE-d optimization (100') method as described can be integrated into various types of PLT (Web-RTC), as well as, generally, into other types of suitable platforms, including in particular: a. Online games: prediction of areas where latency increases are to be expected and adjustment of connection parameters to minimize these effects. b. Video streaming: real-time adaptation of video stream quality in streaming services based on available bandwidth predictions, thus ensuring an optimal user experience even in the event of degraded network conditions. c. Network resource optimization systems and improved bandwidth distribution in high-density user environments.

[0047] The QoE(d) optimization method (100') as described is suitable for various types of network infrastructures, including mobile networks (4G, 5G), satellite networks, etc. It can also be adapted to work with virtual private networks (VPNs), by adjusting its predictions according to the specific characteristics of each infrastructure.

[0048] Apart from transport-related scenarios, the invention can be applied to large-scale environments such as: a. Live events (stadiums, concerts): Management of network saturation problems due to the massive influx of connected users. b. Campus or companies: Anticipation of network usage peaks, for example during peak hours in classrooms or offices.

[0049] The technique according to the invention can be implemented on a reprogrammable computing machine, such as a laptop computer, a DSP processor, or a microcontroller executing a program containing a sequence of instructions. It can also be implemented on a dedicated computing machine, such as an FPGA, an ASIC, or a specialized hardware module.

[0050] In the case of a reprogrammable computing machine, the corresponding program (i.e., the sequence of instructions) can be stored on removable (e.g., an SD card, an external hard drive, or a flash drive) or non-removable storage media. This media must be readable, partially or totally, by a computer or processor.

[0051] Although embodiments have been described with reference to specific examples, it will be evident that various modifications and variations can be made to these embodiments without departing from the general spirit and scope of the system and process described herein. Accordingly, the description and drawings should be considered illustrative rather than restrictive.

[0052] Many modifications and adaptations of the present invention will undoubtedly become obvious to a person skilled in the art after reading the above description. It is understood that the phraseology or terminology used in this document is for descriptive and not limiting purposes. It is also understood that the above description contains numerous specifications, which should not be interpreted as limiting the scope of the invention, but merely as providing illustrations of some of the preferred embodiments of this invention.

[0053] The detailed description includes references to the accompanying drawings, which form an integral part of this description. The drawings illustrate embodiments given by way of example. These embodiments, also referred to herein as "examples," are described in sufficient detail to enable persons competent in the field to implement the present invention. However, it may be obvious to a person with ordinary competence in the field that the present invention can be implemented without these specific details. Furthermore, well-known methods, procedures, and components have not been described in detail so as not to unnecessarily obscure certain aspects of the embodiments.The embodiments may be combined, other embodiments may be used, or structural, logical, and conceptual modifications may be made without departing from the scope of the claims. The detailed description should therefore not be considered limiting, and the scope is defined by the appended claims and their equivalents.

[0054] In this document, the terms "a" or "an", as is common in patent documents, should be understood as including one or more elements. Similarly, the term "or" is used to denote a non-exclusive "or", so that "A or B" includes "A but not B", "B but not A", as well as "A and B", unless otherwise indicated.

Claims

1.

2.

3.

4.

5. Demands QoE-d optimization system (100), characterized in that said system comprises: a. A predictive system (200) for QoS in mobile conditions, for a future route; b. An adaptive system (300) of the connection data flow from the predicted QoS. QoE-d optimization system (100) according to claim 1, wherein the predictive QoS system (200) comprises: a. A “prediction module” (210) producing an initial QoS prediction; b. An "update module" (220) producing an updated QoS prediction. QoE-d optimization system (100) according to claim 1, wherein the adaptive system (300) adapts the connection data stream via an adaptation module allowing the comparison of the updated QoS prediction to fixed or continuous thresholds. QoE-d optimization system (100) according to claim 1, wherein the adaptive system (300) comprises: a. A transition stabilization module (310) allowing the integration of an observation delay (310') preceding the adaptation of the connection data flow; b. A centralized management module (222) allowing coordination of access to the connection (222') based on the predicted QoS; Optimization method (100') for QoE-d, characterized in that said method comprises: a. A QoS prediction phase (200') in mobile conditions, for a future route, established by the predictive system (200); b. An adaptation phase (300') of the connection data flow from the QoS prediction established by the adaptive system (300).

6. A method for optimizing (100') the QoE-d according to claim 5, wherein the QoS prediction step under mobile conditions comprises: a. A planning phase (210'), comprising the following steps: i. Collection of initial data (x(t0)); ii. Analysis; iii. Production of the initial QoS predictions (“PI-QoS(n)” (with n ranging from 0 to N and with a step size ô)); iv. Storage of PI-QoS(n) and x(t0). b. An update phase (220'), comprising the following steps: i. Collection of current data (x(tk)); ii. Updating; iii. Production of updated QoS predictions (PA-QoS(n) (with n ranging from k to N and with a step ô); iv. Storage of PA-QoS(n) and x(tk).

7. A method for optimizing (100') the QoE-d according to claim 5, wherein the step of adapting the connection data flow from the updated QoS prediction comprises the following steps:

8. a. PA-QoS(n) greater than a first threshold "A"? i. YES: the flow of data exchanged remains at maximum: simultaneous transmission of video and audio data. ii. NO: see step 2 b. PA-QoS(n) greater than a second threshold "B"? i. YES: the flow of data exchanged is reduced: the transmission switches to a graphic avatar mode associated with audio. ii. NO: see step 3 c. PA-QoS(n) greater than a third threshold "C"? i. YES: the data exchanged is further reduced: transmission is limited to an audio stream which may be accompanied by an image ii. NO: see step 4 d. PA-QoS(n) greater than a fourth critical threshold "D"? i. YES: the flow of exchanged data is reduced to a minimum: speech transcribed into text for the data sent then reception of the data in text form and / or transcribed into sound form. ii. NO: Connection interruption A method for optimizing (100') the QoE-d according to any one of claims 5 or 7, wherein the step of adapting the connection data stream from the updated QoS prediction includes an observation period (310') preceding the adaptation of the data stream, according to the following steps: a. Detection of a variation in PA-QoS(n) (PA-QoS(n) crosses a threshold); b. Waiting for an observation period (310') of the evolution of the PA-QoS(n); c. Decision to validate or reject the transition: variation maintained during the observation period (310')? i. YES: Transition to the new flow level completed ii. NO (oscillations): transition to the new flux level cancelled.

9. Executable software running on a computer terminal, characterized in that it comprises software modules configured to implement at least one of the following: a. the QoE-d optimization system (100), according to any one of claims 1 to 4; b. the optimization method (100') of the QoE-d, according to any one of claims 5 to 8.

10. Executable software on a computer terminal according to claim 9, characterized in that it comprises software modules configured to implement at least one of the following functions: a. QoS prediction; b. Telecommunications; c. Adaptation of the connection data flow; d. Stabilization of transitions; e. Coordination of access to the connection; f. Scheduling connection slots (recommendation) (212'), reservation (214')); g. Messaging between users, allowing in particular the communication of information on routes.

Images