Electronic system for detecting and managing a malfunction in the flight behavior of a drone, associated method and computer program

Abstract

Electronic system for detecting and managing a malfunction in the flight behavior of a drone, associated method and computer program The invention relates to a system (10) for detecting and managing a malfunction in the flight behavior of a drone, comprising: - an onboard sensor capable of obtaining at least one flight parameter of the drone; - a digital twin (14) of said drone, - a computer (16) configured to predict a virtual flight state of said drone by injecting said at least one flight parameter into the digital twin, - a comparison module (18) configured to: - compare the virtual flight state with a real flight state of said drone, and - provide a state difference, - a detection module (20) configured to compare the state difference with at least one predetermined threshold, and identify the presence / absence of a malfunction. Figure for abstract: Figure 1.

Description

Title of the invention: Electronic system for detecting and managing a malfunction in the flight behavior of a drone, associated method and computer program

[0001] The present invention relates to an electronic system for detecting and managing a malfunction in the flight behavior of a drone.

[0002] The present invention also relates to a method for detecting and managing a malfunction in the flight behavior of a drone, the method being suitable for being implemented by an electronic system for detecting and managing a malfunction in the flight behavior of a drone according to the present invention.

[0003] The invention also relates to a computer program comprising software instructions which, when executed by a computer, implement such a method for detecting and managing a malfunction in the flight behavior of a drone.

[0004] The invention relates to the monitoring of the in-flight behavior of an aircraft (HUMS from the English Health and Usage Monitoring) and in particular of a drone (UAV from the English Unmanned Aerial Vehicles) i.e. a mobile vehicle without a pilot on board, in particular drones, flying at a great distance from the remote pilot (from the English beyond the visital line of sight), for example for inspection flights of linear infrastructures of "great elongation" such as gas pipelines, electric lines, railways, highways.

[0005] The missions carried out by small drones at low altitude, beyond the pilot's view, are typically but not limited to land or maritime surveillance (via on-board cameras / radars / lidars), logistics (transport of parcels).

[0006] Currently, the complexity of the possible flight behaviors of a drone depending on the different state conditions and stimuli makes it difficult to identify abnormal behavior and even more so its origin.

[0007] Indeed, such drones, due to their mass and volume constraints, have a limited degree of internal monitoring of their proper functioning. Indeed, generally drones, to satisfy the aforementioned constraints, are limited in terms of capacity for inserting internal monitoring sensors and redundancy of their systems / constituent elements.

[0008] Unitary monitoring of the good condition or behavior of all elements is also difficult to implement due to the corresponding number of sensors required such as vibration, temperature, shock sensors to indicate degradation trends or accumulations of stresses, especially on drones with constrained mass and volume as indicated above, and / or, to avoid the installation of a dedicated sensor, difficult to implement using an indirect estimation of flight parameters by using flight mechanics equations and measurements of other associated parameters.

[0009] Since drone flight safety is generally prioritized, the rough determination (i.e. with low precision) of a supposedly abnormal real state of the drone leads to excessive precaution by relying, for example, on strain gauges integrated into the structure of the drone and / or on reductions of the mechanical deformation model to deduce conditions of excessive fatigue and / or risk of rupture.

[0010] Such excessive precaution is often penalizing from an operational and economic point of view, and generally results in penalizing conservative decision-making, corresponding for example to an emergency landing and / or to a termination of its mission in the presence, for example, of a wind assumed to be higher than a limit value and / or in the presence, for example, of a speed of progression observed along the trajectory lower than the expected speed of progression.

[0011] Such an untimely termination of the mission is implemented all the more often due to the constraints of size, mass, range associated with drones, which as indicated previously limit the possibility of integrating redundant or oversized elements to generate more possibilities of adaptation / reaction in the face of a supposedly abnormal situation.

[0012] One of the aims of the invention is therefore to propose an electronic system for detecting and managing a malfunction in the flight behavior of a drone capable of knowing the actual operating state in flight of the drone in question, capable of identifying abnormal behavior of the drone, associated with the presence of abnormal external conditions (i.e. beyond the intended area of ​​use), or with the presence of a malfunction of an element of the drone, or even with the degradation of certain of its characteristics, such as its drag, and capable of determining the most appropriate reaction for the flight safety of the drone in question and the effectiveness of its mission.

[0013] To this end, the invention relates to an electronic system for detecting and managing a malfunction in the flight behavior of a drone, the system comprising at least:

[0014] - at least one sensor on board said drone, the drone being configured to collect and transmit at least one current flight parameter of said drone obtained from said at least one sensor;

[0015] - a digital twin of said drone capable of reproducing the flight behavior of said drone,

[0016] - a computer configured to predict a current virtual flight state of said drone in injecting said at least one current flight parameter into said digital twin,

[0017] - a comparison module configured to:

[0018] - compare:

[0019] -the current virtual flight state, obtained from said digital twin, at

[0020] - a current actual flight state of said drone, obtained from said at least one parameter of current flight of said drone, and

[0021] - provide a state gap between the virtual flight state and the real flight state,

[0022] - a module for detecting and managing a malfunction in the com in-flight behavior of said drone configured to compare the state deviation to at least one predetermined threshold, and, depending on the result of the comparison of the state deviation to said at least one predetermined threshold, identify the presence / absence of a type of malfunction in the in-flight behavior of said drone.

[0023] Thus, the present invention consists of using a digital twin of a drone in which at least one flight parameter provided by an onboard sensor is injected in order to compare the results with those provided by the drone in real time to determine an abnormal degradation of the flight behavior of the drone.

[0024] This digital twin corresponds to a digital replica of the drone, for example, accessible by and / or stored in a computer advantageously on the ground to benefit from more power and external information, insofar as the drone is nominally connected, via a two-way communication system, in radio data link with a remote piloting system on the ground (i.e. a control station) so as to receive in real time the flight information(s) / data(s) / parameter(s) provided by said at least one sensor on board the drone, or stored in a computer on board the drone. Such a digital twin is supplied by the flight parameter(s) captured in real time by the drone in order to virtually produce the state (i.e. the behavior) expected of the drone in the absence of disturbance(s) / degradation(s).

[0025] Comparing the virtual state provided by the digital twin and the real state provided directly by the drone makes it possible to reliably determine the presence or absence of a malfunction in the flight behavior of the drone in question.

[0026] According to particular embodiments, the electronic system for detecting and managing a malfunction in the flight behavior of a drone comprises one or more of the following characteristics, taken in isolation or in all technically possible combinations:

[0027] - said digital twin of said drone comprises at least one digital model of the internal design of said drone, each digital model being associated with a predetermined group of flight parameter(s), and / or each digital model being customizable by taking into account the type of drone and / or the serial number of said drone, the serial number being representative of the operating history of said drone since its commissioning;

[0028] - each digital model is suitable for being enriched on each flight using the data collected during the previous flight of said drone;

[0029] - the system further comprises an electronic analysis module configured to analyzing the actual flight state, when the state deviation is greater than said at least one predetermined threshold, by determining at least one possible source of the malfunction associated with said state deviation;

[0030] - the electronic analysis module is configured to execute at least one element belonging to the group comprising at least:

[0031] - a logic tree capable of associating the state gap with said at least one source possible,

[0032] - a set of predetermined analysis rules;

[0033] - a neural network previously trained by learning the data collected during previous flights of said drone;

[0034] - a combination of said preceding elements;

[0035] - said state difference is one-dimensional or multidimensional, and corresponds to:

[0036] - at least one difference between two virtual and real state values;

[0037] - a level of dispersion between the virtual state and the real state;

[0038] - a correlation rate between the virtual state and the real state.

[0039] - the electronic analysis module is further configured to associate a pro reliability at each possible source of the malfunction associated with said state difference;

[0040] - the electronic analysis module is further configured to provide, depending on the type of malfunction identified:

[0041] - at least one pilot action suggestion and / or at least one command of piloting, to the remote pilot of said drone or directly to an onboard control system of said drone, and / or

[0042] - at least one maintenance action suggestion.

[0043] The invention also relates to a method for detecting and managing a malfunction in the flight behavior of a drone, the method being suitable for being implemented by an electronic system for detecting and managing a malfunction in the flight behavior of a drone as mentioned above, the method comprising the following steps:

[0044] - prediction of a current virtual flight state of said drone by injecting, within a digital twin of said drone capable of reproducing the flight behavior of said drone, at least one current flight parameter collected and transmitted by said drone, said at least one flight parameter being obtained from at least one sensor on board said drone,

[0045] - comparison of:

[0046] -the current virtual flight state, obtained from said digital twin, at

[0047] - a current actual flight state of said drone, obtained from said at least one parameter of current flight of said drone,

[0048] - providing a state gap between the virtual flight state and the real flight state,

[0049] - detection and management of a malfunction in the flight behavior of said drone by comparing the state deviation to at least one predetermined threshold, and, depending on the result of the comparison of the state deviation to said at least one predetermined threshold, by identifying the presence / absence of a type of malfunction in the flight behavior of said drone.

[0050] The invention also relates to a computer program comprising software instructions which, when executed by a computer, implement a method for detecting and managing a malfunction in the flight behavior of a drone as defined above.

[0051] Hereinafter, reference to a computer program which, when executed, performs any of the foregoing software instructions, is not limited to an application program running on a single host computer.

[0052] In other words, the terms computer program and software are used hereinafter in a general sense to refer to any type of computer code (e.g., application software, firmware, microcode, or any other form of computer instruction) that can be used to program one or more processors to implement the method of detecting and managing a malfunction in the flight behavior of a drone as defined above.

[0053] These characteristics and advantages of the invention will appear more clearly on reading the description which follows, given solely by way of non-limiting example, and made with reference to the appended drawings, in which:

[0054] - [fig.l] is a schematic representation of an electronic detection system and managing a malfunction in the flight behavior of a drone according to the present invention;

[0055] - [fig.2] is a flowchart of a method for detecting and managing a dys operation in the flight behavior of a drone according to the invention.

[0056] Conventionally in the present application, the expressions “substantially equal to” and “approximately” will each express a relationship of equality to plus or minus 10%.

[0057] In the example of [fig.l], the electronic system 10 for detecting and managing a malfunction in the flight behavior of a drone, not shown, according to the present invention firstly comprises an assembly 12 on board this drone, this assembly comprising at least one sensor such as at least one sensor of speed(s), a location device such as a GNSS (Global Navigation Satellite System) satellite positioning device, a propulsion power sensor.

[0058] The assembly 12 is particularly suitable for measuring the speed and torque on a fixed-pitch propeller shaft, or more directly, for electric motors, for measuring the intensity of the current and the supply voltage combined with the temperature and the external pressure, or alternatively, for jet motors, for measuring a force on the reactor mast for the thrust, in order to obtain a good estimate of the propulsive force data.

[0059] The parameters provided by these sensors make it possible to construct a real state vector of the drone.

[0060] The drone is configured to collect and transmit, within said system 10, at least one current flight parameter of the drone obtained from said at least one sensor of the assembly 12. As an optional addition, the assembly 12 also comprises at least one flight control computer on board the device, the drone then also being configured to collect and transmit, within said system 10, the current orders to the motors and control surfaces of the drone for example.

[0061] Furthermore, according to the present invention, the electronic system 10 for detecting and managing a malfunction in the flight behavior of a drone comprises a digital twin 14 of the drone in question, such a digital twin 14 being capable of reproducing the flight behavior of said drone.

[0062] By "digital twin", is meant hereinafter according to the present invention a digital replica 14 of the drone in question. Such a digital replica is accessible by and / or stored in a ground computer insofar as the drone is nominally connected in real time, via a bidirectional communication system on board the drone, in radio data link with a ground remote piloting system so as to receive in real time the flight information(s) / data(s) / parameter(s) provided by said at least one assembly 12 on board the drone and / or by one or more flight computers on board the drone, or corresponds directly to another twin aircraft of the drone in question.

[0063] Such a digital replica (i.e. digital twin) 14 is personalized or not by means of the type of drone (i.e. device) and / or the serial number of the drone considered in order to reproduce its behavior as faithfully as possible.

[0064] According to a particular optional aspect of the present invention, the digital twin 14 of the drone comprises at least one digital model of the internal design of the drone in question, each digital model being associated with a predetermined group of flight parameter(s), and / or each digital model being customizable by taking into account the serial number of the drone, the serial number being representative of the operating history of the drone since it was put into service and in particular any damage / wear specific to this serial number.

[0065] In particular, each digital model of the drone defined during its design (i.e. its development, its study with a view to development) is capable of modeling the behavior of the drone, in particular in terms of aerodynamics and / or performance and / or redundancies.

[0066] For example, the digital twin 14 of the drone comprises a digital model of “Average Performance” based on the use of sensor data from the set 12 corresponding to the energy consumption as a function of the initial mass of the drone, the distance traveled, the variation of kinetic and potential energy, the speed, the wind / temperature / pressure conditions along the path associated with the mission of the drone. Such a digital model of “Average Performance” is in particular suitable for indicating a degradation of the aerodynamic profile of the drone corresponding to a mechanical deformation, icing, etc., or for indicating a degradation of the engine efficiency or of a propeller of the drone.

[0067] As an optional addition, the digital twin 14 of the drone comprises a digital “Dynamic” model based on the use of sensor data from the set 12 corresponding to angular speeds and linear accelerations, measured by the sensors of the set 12 corresponding to gyrometers and / or accelerometers on board the drone, and this as a function of the propulsive forces, and / or corresponding to controls on the control surfaces, to the air speed vector measured by the sensors of the set 12 corresponding to the pitot tube for the air speed vector module and to a wind vane for the incidence and sideslip of this air speed vector, and / or corresponding to the wind and the air density by measuring the ground speed provided by a GNSS receiver (for Geolocation and Navigation by Satellite System) of the set 12 on board the drone, measurement of the magnetic heading, measurement of the barometric pressure, measurement of the outside temperature, etc.Such a numerical “Dynamic” model is particularly suitable for indicating a lack of efficiency of the control surfaces or engines, excessive turbulence, wind shear, abnormal vibrations of the structure or engines, etc.

[0068] As an optional addition, the digital twin 14 of the drone comprises a digital “Static Equilibrium” model based on the use of sensor data from the assembly 12 corresponding to a speed and a slope of the instantaneous trajectory, and this as a function of the propulsive force, corresponding to controls on the control surfaces, corresponding to the wind, corresponding to pressure and to external temperatures. Such a digital “Static Equilibrium” model is in particular suitable for indicating a degradation of the aerodynamic profile of the drone corresponding to a mechanical deformation, to icing, etc., or even for indicating a degradation of the engine or propeller performance of the drone, or to indicate a degradation of the efficiency of the control surfaces.

[0069] Other known digital models of drone behavior are also suitable for use as an alternative and / or in addition.

[0070] In addition, the electronic system 10 for detecting and managing a malfunction in the flight behavior of a drone according to the present invention also comprises a computer 16 configured to predict a current virtual flight state of said drone by injecting said at least one current flight parameter within (i.e. as input) said digital twin 14, and / or by injecting as indicated previously the information provided by said at least one flight control computer on board the drone such as the current orders to the motors and control surfaces of the drone.

[0071] A virtual and / or real flight state comprises at least one position and / or one attitude of the drone as well as its first derivatives (i.e. speeds, etc.) and second derivatives (i.e. accelerations), and / or its mass including the quantity of fuel on board, and / or its battery level, and / or its engine speeds, and / or its control surface deflections, etc.

[0072] Furthermore, the electronic system 10 for detecting and managing a malfunction in the flight behavior of a drone according to the present invention also comprises a comparison module 18 configured to compare the current virtual flight state, obtained from said digital twin 14, with a current real flight state of the drone, obtained from said at least one current flight parameter of said drone, and configured to provide a state difference between the virtual flight state and the real flight state.

[0073] Furthermore, the electronic system 10 for detecting and managing a malfunction in the flight behavior of a drone according to the present invention also comprises a module 20 for detecting and managing a malfunction in the flight behavior of said drone configured to compare the deviation of states to at least one predetermined threshold, and, depending on the result of the comparison of the deviation of states to said at least one predetermined threshold, identify the presence / absence of a type of malfunction in the flight behavior of said drone.

[0074] As an optional addition, the module 20 for detecting and managing a malfunction in the flight behavior of said drone is also configured as illustrated by the arrow 21, to transmit the processed data to a database Bp of previously collected data.

[0075] As an optional addition, the electronic system 10 for detecting and managing a malfunction in the flight behavior of a drone also comprises a module 22 for refining (i.e. enriching) the digital twin 14. More precisely, via such an optional refinement module 22, each digital model of the digital twin 14 is capable of being enriched in a recurring manner during each flight in using the data collected during the previous flight of the drone and stored within the previously collected database, in relation to the current moment of the current mission of the drone in question.

[0076] Thus, according to this optional aspect, the digital twin 14 is enriched / refined, via the optional refinement module 22, during previous flights, test flights, or even potentially throughout the life of the device.

[0077] In particular, notably to integrate wear and tear capable of reducing the performance of the drone in question, when the difference in states between the virtual flight state and the real flight state determined by the module 20 for detecting and managing a malfunction in the flight behavior of said drone is minimal, i.e. less than a predetermined threshold, and therefore representative of “normal” flight behavior, such a difference is capable of being stored in the database of previously collected data Bp to be reinjected, via the optional refinement module 22, into the digital twin 14 to improve the internal design digital model(s) of the drone.

[0078] Furthermore, in a manner not shown, the optional refinement module 22 is also capable of more global environmental data received from sources external to the drone, for example an online accessible source not shown capable of characterizing the winds, pressures, temperatures present in the geographical area and in the time slot associated with the current mission of the drone, in addition to the flight parameters provided via the sensors installed on the drone.

[0079] Indeed, to eliminate erroneous sensor measurements due to malfunction of one or more on-board sensors of the assembly 12 and likely to distort the malfunction detection subsequently implemented by the module 20 for detecting and managing a malfunction in the flight behavior of said drone, maximum redundancy of the data collected and supplied to the digital twin 14 is sought. Thus, the system 10 proposed according to the present invention has the capacity to differentiate an abnormal situation likely to be dangerous from a situation associated with a sensor fault with no consequence on flight safety.

[0080] As an optional addition, the electronic system 10 for detecting and managing a malfunction in the flight behavior of a drone also comprises an electronic analysis module 24 configured to analyze the actual flight state, when the difference in states (virtual and actual) is greater than said at least one predetermined threshold, by determining at least one possible source of the malfunction associated with said difference in states.

[0081] In other words, the electronic analysis module 24 is capable of assessing the difference in states by comparing it to at least one predetermined threshold.

[0082] According to a particular optional aspect, the state difference, determined and supplied to the electronic analysis module 24 by the comparison module 18, via the detection module 20, is one-dimensional or multidimensional, and corresponds to:

[0083] - at least one (raw) difference between two virtual and real state values;

[0084] - a level of dispersion between the virtual state and the real state (eg a standard deviation at the average) ;

[0085] - a correlation rate between the virtual state and the real state (i.e. a correlation rate of a sequence of virtual and real state values).

[0086] For example, the state deviation corresponds to a slow drift of a parameter such as an increasing lateral deviation between the real position of the drone and the position estimated by the digital twin 14, a sign of a possible drift of position sensors (i.e. a location device such as a GNSS (Geographic and Navigation by Satellite System) satellite positioning device), and / or corresponds to an increasing deviation between the real vertical trajectory of the drone and that of its digital twin 14, a sign of a possible lift problem, linked either to the available power of the real drone, or to unfavorable and constant meteorological conditions, for example a headwind.

[0087] Such an electronic analysis module 24 corresponds in particular to a root cause analysis and decision-making module, itself predefined by use, during the design phase of the drone considered, of the digital model(s) of its behavior, to test the different situations and influencing factors.

[0088] According to a particular aspect of this optional supplement, the electronic analysis module 24 is configured to execute at least one element belonging to the group comprising at least:

[0089] - a logic tree capable of associating the state gap with said at least one source possible,

[0090] - a set of predetermined analysis rules;

[0091] - a neural network previously trained by learning the data collected during previous flights of said drone;

[0092] - a combination of said preceding elements.

[0093] In particular, the set of predetermined analysis rules is notably obtained by using mathematical formulas from the design phase of the drone making it possible to trace back to a possible source of malfunction from the observed state difference.

[0094] According to another particular aspect of this optional supplement, the electronic analysis module 24 is further configured to associate a probability with each possible source of the malfunction associated with the state deviation. In particular, in the event of uncertainties about the source of malfunction, each possible source of the malfunction operation is provided to the remote pilot and / or to a maintenance platform with their characteristics in terms of probability or tolerance intervals.

[0095] In particular, the electronic analysis module 24 is capable of associating a confidence index with the detection result provided by the module 20 for detecting and managing malfunctions in the flight behavior of the drone.

[0096] Such a confidence index is for example expressed in terms of likelihood probability of a binary event, or in terms of confidence interval for a given validity probability on a continuous quantity. This confidence index exploits the observed level of redundancy of the data used to obtain the virtual state and the detection result provided by the detection module 20, the level of quality of the influential data from the reliability and precision associated with each sensor, the quality of the digital model(s) of the digital twin 14 and the operating history of the drone taken into account from the database collected beforehand Bp.

[0097] For example, in the presence of a malfunction (i.e. corresponding to a difference in states whose value is greater than at least one predetermined threshold), the electronic analysis module 24 indicates, by display and / or by sound broadcast and / or by transmission of a message, to the remote pilot and / or to a maintenance platform, that the malfunction is 80% probably due to severe local turbulence (i.e. external to the drone) and 20% probably due to an accelerometer whose measurement along a displacement axis, for example a y axis not shown, is defective). In other words, according to this example, the electronic analysis module 24 indicates that the malfunction has an 80% chance of being due to local turbulence and a 20% chance of being due to a defective accelerometer (i.e. sensor) on board the aircraft.

[0098] According to another particular aspect of this optional addition, the electronic analysis module 24 is further configured to provide, via a decision-making tool 26, integrated into the electronic analysis module 24 in a manner not shown, or outside the electronic analysis module 24 as illustrated by [fig.l], depending on the type of malfunction identified:

[0099] - at least one pilot action suggestion and / or at least one command of piloting, to the remote pilot of said drone or directly to an onboard control system of said drone (i.e. automatic pilot), and / or

[0100] - at least one maintenance action suggestion.

[0101] More specifically, the decision-making tool 26 provides recommendations or reaction orders, sent either to the pilot or to the flight control computers on board the drone, in particular in the case of an automated reaction.

[0102] Possible recommended actions are, for example, a check of the distance that can be covered from the remaining fuel or energy and the observed consumption, and / or a more or less significant reconfiguration of the flight, going from a change of heading and / or altitude and / or speed to a descent and an emergency landing, and / or a reconfiguration of the drone, for example, by switching to a backup battery, to a backup motor, to an alternative flight control mode such as the use of vector thrust instead of or in addition to the moving surfaces, and / or maintenance to be carried out after the flight of the drone, before a predetermined time interval, for example, to change the battery, the motor, check the control of the rudder along an x ​​axis not shown, or even a defective sensor such as the aforementioned accelerometer.

[0103] These actions can themselves be recorded, their effects subsequently noted, via the sensors of the set 12 or human annotations, to be optionally used, as illustrated by the arrow 28, to enrich the analysis module 24.

[0104] According to a particular aspect, two predetermined thresholds are used successively, according to the present invention, by the electronic analysis module 24 to indicate, via the decision-making tool 26, the action to be implemented, namely a first vigilance threshold and a second alert threshold. If the state difference assessed by the analysis module 24 is greater than the first vigilance threshold, the decision-making tool 26 recommends / triggers a first reconfiguration of the drone's mission, in the form of a first control datum, and if the state difference is greater than the second alert threshold, the decision-making tool 26 recommends / triggers a second reconfiguration of the drone's mission, in the form of a second control datum.

[0105] The first or second control data is transmitted to the drone via the two-way communication system, optionally passing beforehand through the (remote)pilot capable of validating it, ignoring it, translating it into other elementary actions and orders, the drone executing in response the control data via its flight control computer which controls the drone's motors, actuators, selectors, etc.

[0106] In the example of [fig.l], the electronic system 10 for detecting and managing a malfunction in the flight behavior of a drone comprises at least one information processing unit (not shown) formed for example of a memory and a processor associated with the memory, both not shown.

[0107] In the example of [fig.l], the digital twin 14, the calculator 16, the comparison module 18 and the module 20 for detecting and managing a malfunction in the flight behavior of said drone, and optionally the refinement module 22 as well as the analysis module 24, the decision-making tool 26, are each produced in the form of software, or a software brick, executable by at least one processor of the electronic system 10 for detecting and managing a malfunction operation in the flight behavior of a drone. At least one memory of the electronic system 10 for detecting and managing a malfunction in the flight behavior of a drone is then capable of storing at least one software corresponding to each of these elements 14, 16, 18 and 20, and optionally 22, 24 and 26. At least one processor of the electronic system 10 is then capable of executing at least one or each of these software programs.

[0108] Indeed, according to a first variant, the electronic system 10 for detecting and managing a malfunction in the flight behavior of a drone comprises a single information processing unit whose memory is capable of storing each software corresponding to each of the elements 14, 16, 18 and 20, and optionally 22, 24 and 26 and whose processor is then capable of executing each of these software programs.

[0109] According to a second variant, the electronic system 10 for detecting and managing a malfunction in the flight behavior of a drone is distributed in distinct (i.e. separate) devices each associated respectively with the digital twin 14, the computer 16, and the comparison module 18 and the module 20 for detecting and managing a malfunction in the flight behavior of said drone, and optionally the refinement module 22 as well as the analysis module 24 and the decision-making tool 26. Each device comprising an information processing unit whose memory is capable of storing software from among those associated with the elements 14, 16, 18 and 20, and optionally 22, 24 and 26 and whose processor is capable of executing this software.

[0110] According to a first example of this second variant, the calculator 16 is integrated within a processing unit located on the ground outside the drone and / or the comparison module 18 and / or the module 20 for detecting and managing a malfunction in the flight behavior of said drone, and / or optionally the refinement module 22 and / or optionally the analysis module 24 and the decision-making tool 26, is integrated within another processing unit located on the ground outside the drone. The architecture in accordance with this first example thus aims to bring down to the ground the few parameters observed by the drone and then to send back the orders accordingly, and allows real-time use of a computing capacity remote to the ground, the drone sending its own sensor information via a downlink for malfunction detection and, where appropriate, decision-making on the ground, as an alternative to an on-board calculation based on global uplink information.

[0111] According to a second example of this second variant, the calculator 16 is integrated within a processing unit on board the drone and / or the comparison module 18 is integrated within another processing unit on board the drone and / or the module 20 for detecting and managing a malfunction in the flight behavior of said drone is integrated within another processing unit on board the drone and / or optionally the refinement module 22 is integrated within another processing unit on board the drone and / or optionally the analysis module 24 and the decision-making tool 26 is integrated within another processing unit on board the drone, which makes it possible to keep the detection of a malfunction on board the drone, for example for cases of loss of communication with the ground (i.e. cases / phases where the control / command data link of the drone is cut), in order to still be able to take conservative automatic decisions in the presence of a malfunction.

[0112] According to a third variant, at least two elements among the software associated with the digital twin 14, the computer 16, and the comparison module 18 and the module 20 for detecting and managing a malfunction in the flight behavior of said drone, and / or optionally the refinement module 22 and / or optionally the analysis module 24 and the decision-making tool 26, are associated with the same processing unit.

[0113] For example, the calculator 16 and the comparison module 18 are integrated within the same processing unit.

[0114] According to another variant, the comparison module 18 is for example directly integrated into the computer 16 produced in the form of a single software program, or a software brick.

[0115] In a variant not shown, the digital twin 14, the computer 16, the comparison module 18 and the module 20 for detecting and managing a malfunction in the flight behavior of said drone, and / or optionally the refinement module 22 and / or optionally the analysis module 24 and the decision-making tool 26, are each produced in the form of a programmable logic component, such as an FPGA (Field Programmable Gate Array) or in the form of a dedicated integrated circuit, such as an ASIC (Application Specific Integrated Circuit).

[0116] When the electronic system 10 for detecting and managing a malfunction in the flight behavior of a drone is produced in the form of one or more software programs, that is to say in the form of a computer program, it is also capable of being recorded on a medium, not shown, readable by a computer. The computer-readable medium is for example a medium capable of storing electronic instructions and of being coupled to a bus of a computer system. By way of example, the readable medium is a ROM memory or a RAM memory, any type of non-volatile memory for example EPROM, EEPROM, FLASH, NVRAM, etc. A computer program comprising software instructions is then stored on the readable medium.

[0117] The operation of the electronic system 10 for detecting and managing a malfunction in the flight behavior of a drone according to the invention will now be described with the aid of [fig.2] illustrating a flowchart of an exemplary embodiment of the method 30 for detecting and managing a malfunction in the flight behavior of a drone according to the invention, implemented by computer.

[0118] According to a first step 32, the electronic system 10 for detecting and managing a malfunction in the flight behavior of a drone, via the computer 16, implements a prediction 32 of a current virtual flight state of said drone by injecting, within the digital twin 14 of the drone, the digital twin 14 being capable of reproducing the flight behavior of the drone, at least one current flight parameter collected and transmitted by said drone, said at least one flight parameter being obtained from at least one sensor of the assembly 12 on board the drone.

[0119] Then, according to a step 34, the electronic system 10 for detecting and managing a malfunction in the flight behavior of a drone, via the comparison module 18, implements the comparison of the current virtual flight state, obtained from the digital twin 14, with a current real flight state of the drone, obtained from said at least one current flight parameter of the drone.

[0120] Then according to a step 36, the electronic system 10 for detecting and managing a malfunction in the flight behavior of a drone, via the comparison module 18, implements the provision of a difference in states between the virtual flight state and the real flight state to the module 20 for detecting and managing a malfunction in the flight behavior of said drone.

[0121] Finally, during a step 37, the electronic system 10 for detecting and managing a malfunction in the flight behavior of a drone, via the module 20 for detecting and managing a malfunction in the flight behavior of said drone, implements the detection and management of a malfunction in the flight behavior of said drone by comparing the difference in states to at least one predetermined threshold, and, depending on the result of the comparison of the difference in states with said at least one predetermined threshold, by identifying the presence / absence of a type of malfunction in the flight behavior of said drone.

[0122] Those skilled in the art will understand that the invention is not limited to the embodiments described, nor to the particular examples of the description, the embodiments and variants mentioned above being suitable for being combined with each other to generate new embodiments of the invention.

[0123] It is thus understood that the electronic system 10 and the method 30 for detecting and managing a malfunction in the flight behavior of a drone proposes to use one or more sensors on board the drone to measure the main parameters of the flight and the environment of the drone in question, including the data are injected into a digital twin of the aircraft including flight mechanics and performance models, optionally enriched with observations accumulated on previous flights.

[0124] The present invention therefore proposes, by means of a computer, which advantageously can be on the ground to benefit from more power and external information, to use a digital twin of the type of drone, where appropriate personalized by serial number, to detect a malfunction in real time and reliably while having the capacity to differentiate an abnormal situation likely to be dangerous from a situation associated with a sensor fault with no consequences for flight safety.

[0125] Thus, the present invention makes it possible to avoid and / or limit the use of sensors dedicated to monitoring the structure of the drone or the moving control surfaces.

Claims

Claims

1. Electronic system (10) for detecting and managing a malfunction in the flight behavior of a drone, the system being characterized in that it comprises at least: - at least one sensor on board said drone, the drone being configured to collect and transmit at least one current flight parameter of said drone obtained from said at least one sensor;- a digital twin (14) of said drone capable of reproducing the flight behavior of said drone, - a computer (16) configured to predict a current virtual flight state of said drone by injecting said at least one current flight parameter into said digital twin, - a comparison module (18) configured to: - compare: - the current virtual flight state, obtained from said digital twin, with - a current real flight state of said drone, obtained from said at least one current flight parameter of said drone, and - provide a state difference between the virtual flight state and the real flight state, - a module (20) for detecting and managing a malfunction in the flight behavior of said drone configured to compare the state difference to at least one predetermined threshold, and, depending on the result of the comparison of the state difference to said at least one predetermined threshold, identify the presence / absence of a type of malfunction in the flight behavior of said drone.;

2. An electronic system (10) according to claim 1, wherein said digital twin (14) of said drone comprises at least one digital model of the internal design of said drone, each digital model being associated with a predetermined group of flight parameter(s), and / or each digital model being customizable by taking into account the type of drone and / or the serial number of said drone, the serial number being representative of the operating history of said drone since it was put into service.

3. Electronic system (10) according to claim 2, in which each digital model is capable of being enriched at each flight using the data collected during the previous flight of said drone.

4. An electronic system (10) according to any preceding claim, further comprising an electronic module (24) analysis configured to analyze the actual flight state, when the state deviation is greater than said at least one predetermined threshold, by determining at least one possible source of the malfunction associated with said state deviation.

5. Electronic system (10) according to claim 4, in which the electronic analysis module (24) is configured to execute at least one element belonging to the group comprising at least: - a logic tree capable of associating the state difference with said at least one possible source, - a set of predetermined analysis rules; - a neural network previously trained by learning data collected during previous flights of said drone; - a combination of said previous elements.

6. Electronic system (10) according to any one of the preceding claims, in which said state difference is one-dimensional or multidimensional, and corresponds to: - at least one difference between two virtual and real state values; - a level of dispersion between the virtual state and the real state; - a correlation rate between the virtual state and the real state.

7. Electronic system (10) according to any one of claims 4 to 6, in which the electronic analysis module (24) is further configured to associate a probability with each possible source of the malfunction associated with said state deviation.

8. Electronic system (10) according to any one of claims 4 to 7, in which the electronic analysis module (24) is further configured to provide, depending on the type of malfunction identified: - at least one piloting action suggestion and / or at least one piloting command, to the remote pilot of said drone or directly to an onboard control system of said drone, and / or - at least one maintenance action suggestion.

9. Method (30) for detecting and managing a malfunction in the flight behavior of a drone, the method being suitable for being implemented by an electronic system (10) for detecting and managing a malfunction in the flight behavior of a drone according to any one of the claims, the method comprising the following steps: - prediction (32) of a current virtual flight state of said drone in injecting, within a digital twin of said drone capable of reproducing the flight behavior of said drone, at least one current flight parameter collected and transmitted by said drone, said at least one flight parameter being obtained from at least one sensor on board said drone, - comparison (34) of: - the current virtual flight state, obtained from said digital twin, to - a current real flight state of said drone, obtained from said at least one current flight parameter of said drone, - providing (36) a state difference between the virtual flight state and the real flight state, - detection and management (37) of a malfunction in the flight behavior of said drone by comparing the state deviation to at least one predetermined threshold, and, depending on the result of the comparison of the state deviation to said at least one predetermined threshold, by identifying the presence / absence of a type of malfunction in the flight behavior of said drone.

10. Computer program comprising software instructions which, when executed by a computer, implement a method for detecting and managing a malfunction in the flight behavior of a drone according to the preceding claim.

Images