Method and system for anemometric parameter estimation for aircraft using trained neural networks based on wing geometry and flight domain

Abstract

There are disclosed a system and method for estimating an air data parameter for an aircraft. The method comprises obtaining an estimation of a Mach regime for the aircraft; obtaining an indication of at least one secondary surface configuration of the aircraft; selecting a trained neural network based on the obtained estimation of the Mach regime and the indication of at least one secondary surface configuration; obtaining additional data suitable for estimating the air data parameter for the aircraft; accessing the selected trained neural network with at least the obtained additional data suitable for estimating the air data parameter for the aircraft; obtaining, using the selected trained neural network based on the Mach regime, the indication of the at least one secondary surface configuration and the additional data, an estimation of the air data parameter and providing the estimation of the air data parameter.

Description

METHOD AND SYSTEM FOR ANEMOMETRIC PARAMETER ESTIMATION FOR AIRCRAFT USING TRAINED NEURAL NETWORKS BASED ON WING GEOMETRY AND FLIGHT DOMAINCROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This application claims priority of US Provisional Patent Application No.63 / 710,988, entitled “Method and system for anemometric parameter estimation for aircraft using trained neural networks based on wing geometry and flight domain" that was filed on October 23, 2024, 2024, the specification of which is hereby incorporated by reference.FIELD

[0002] The present technology pertains to the field of aeronautics. More precisely, the present technology relates to a method and a system for anemometric parameter estimation for aircraft using trained neural networks based on hydraulic pressure, wing geometry and flight domain.BACKGROUND

[0003] Being able to estimate an air data parameter for an aircraft is of great advantage.

[0004] A solution, disclosed in CA 3,055,490, has been developed and uses the various hydraulic pressure sensors on the primary surface actuators, coupled to the primary and secondary surface position, to the altitude and the inertial reference parameters, to estimate the air data parameters (i.e. True Airspeed (TAS), Calibrated Airspeed (CAS), Mach number (Mach), Angle of Attack (AoA) and Sideslip Angle (AoS)) for an aircraft.

[0005] Unfortunately, the solution suffers from a couple of limitations.

[0006] First of all, the modeling of the airflow on the aircraft wing can be complex and depends on the airspeed, as well as the geometry of the wing.

[0007] Indeed, non-linearities appear in the modeling of the airflow on the wings, when the aircraft is close to stall or when it is close to the speed of sound (i.e. in the transonic zone).

[0008] Moreover, the modification of the wings’ geometry of an aircraft, due to the displacement of the flap and slat surfaces, also complicates the modeling of the airflow.

[0009] By increasing the size of the neural network, i.e., its number of layers or its number of neurons per layer, it is possible to integrate these different airflow modeling.

[0010] Unfortunately, this is not a viable solution.

[0011] In fact, the various flight computers of an airplane (Line Remote Unit - LRU) must perform a large number of complex operations in real time, and moreover, have availability constraints, which mean that the computation time of an LRU is limited and has to be controlled. This is why the integration of a neural network into an LRU is complex, given the number of operations that it realizes.

[0012] There is a need for at least one of a method and a system that will overcome at least one of the above-identified drawbacks.SUMMARY

[0013] It is an object of the present technology to ameliorate at least some of the inconveniences present in the prior art. One or more implementations of the present technology may provide and / or broaden the scope of approaches to and / or methods of achieving the aims and objects of the present technology.

[0014] Thus, one or more implementations of the present technology are directed to a method for estimating an air data parameter for an aircraft.

[0015] There is on one hand a need to increase the size of the neural network, to take into account the complexity of the airflow modelization, and on the other hand, the need to reduce the size of the neural network to facilitate its integration into an LRU.

[0016] According to one or more aspects of the technology, there is disclosed a computer-implemented method for estimating an air data parameter for an aircraft, the method being executed by at least one processor, the method comprising obtaining an estimation of a Mach regime for the aircraft; obtaining an indication of at least one secondary surface configuration of the aircraft; selecting a trained neural network based on the obtained estimation of the Mach regime and the indication of at least one secondary surface configuration; obtaining additional data suitable for estimating the air data parameter for the aircraft; accessing the selected trained neural network with at least the obtained additional data suitable for estimating the air data parameter for the aircraft; obtaining, using the selected trained neural network based on the Mach regime, the indication of the at least one secondary surface configuration and the additional data, an estimation of the air data parameter; and providing the estimation of the air data parameter.

[0017] In accordance with one or more implementations, the obtaining of the additional data suitable for estimating the air data parameter for the aircraft comprises obtaining an indication of a position of a horizontal primary surface of the aircraft; obtaining an estimation of a load exerted on a respective actuator for actuating the horizontal primary surface of the aircraft using at least the indication of the position of the horizontal primary surface and a position of the respective actuator; wherein the selected trained neural network is accessed with an indication of the at least one secondary surface configuration, the estimation of a load exerted on the respective actuator for actuating the horizontal primary surface of the aircraft, and the indication of the position of the horizontal primary surface of the aircraft.

[0018] In accordance with one or more implementations, the air data parameter is selected from a group consisting of a true airspeed (TAS), a calibrated airspeed (CAS), a Mach number, an angle of attack (AoA) and a slide slip angle (AoS).

[0019] In accordance with one or more implementations, the Mach regime is one of a stall model, a subsonic model, a transonic model and a supersonic model.

[0020] In accordance with one or more implementations, the selecting the trained neural network based on the obtained estimation of the Mach regime and the indicationof at least one secondary surface configuration comprises selecting the trained neural network from a set of trained neural networks, each trained neural network being configured to estimate the air data parameter according to at least a respective Mach regime, a respective range of secondary surface configurations and a respective range of additional data suitable for estimating the air data parameter for the aircraft.

[0021] In accordance with one or more implementations, each trained neural network has a same architecture.

[0022] In accordance with one or more implementations, the at least one processor is part of a line remote unit (LRU).

[0023] According to one or more aspects of the technology, there is disclosed a computer-implemented method for training a set of neural networks to estimate an air data parameter for an aircraft, the method being executed by at least one processor, the method comprising obtaining, for each neural network of the set of neural networks, a respective training data set comprising a plurality of training examples, each training example comprising: a respective Mach Regime, a respective indication of at least one secondary surface configuration of the aircraft, respective additional data suitable for estimating the air data parameter for the aircraft, and a respective air data parameter for the aircraft; training each neural network of the set of neural networks on the respective training data set to determine the respective air data parameter based on the respective Mach regime, the respective indication of at least one secondary surface configuration of the aircraft, and the respective additional data suitable for estimating the air data parameter for the aircraft; and outputting a set of trained neural networks, each trained neural network being configured to estimate the air data parameter of the aircraft according to a respective Mach Regime,

[0024] In accordance with one or more implementations, each trained neural network of the set of trained neural networks has a same architecture.

[0025] According to one or more aspects of the technology, there is disclosed a system for estimating an air data parameter for an aircraft, the system comprising: one or more processing devices configured for: obtaining an estimation of a Mach regime;obtaining an indication of at least one secondary surface configuration; selecting a trained neural network based on the obtained estimation of a Mach regime and the indication of at least one secondary surface configuration; obtaining additional data suitable for estimating the air data parameter for the aircraft; accessing the selected trained neural network with at least the obtained additional data suitable for estimating the air data parameter for the aircraft; obtaining an estimation of the air data parameter; and providing the estimation of the air data parameter.

[0026] According to one or more aspects of the technology, there is disclosed a processing device comprising a central processing unit; a display device; a communication port; a memory unit comprising an application for estimating an air data parameter for an aircraft, the application comprising instructions for: obtaining an estimation of a Mach regime; obtaining an indication of at least one secondary surface configuration; selecting a trained neural network based on the obtained estimation of a Mach regime and the indication of at least one secondary surface configuration; obtaining additional data suitable for estimating the air data parameter for the aircraft; accessing the selected trained neural network with at least the obtained additional data suitable for estimating the air data parameter for the aircraft; obtaining an estimation of the air data parameter; and providing the estimation of the air data parameter.

[0027] According to one or more implementations of the technology, there is disclosed a non-transitory machine-readable medium carrying instructions, which, when executed by at least one processor, perform one or more implementations of the methods disclosed above.Terms and Definitions

[0028] In the context of the present specification, a “server” is a computer program that is running on appropriate hardware and is capable of receiving requests (e.g., from electronic devices) over a network (e.g., a communication network), and carrying out those requests, or causing those requests to be carried out. The hardware may be one physical computer or one physical computer system, but neither is required to be the case with respect to the present technology. In the present context, the use of the expression “a server” is not intended to mean that every task (e.g., received instructionsor requests) or any particular task will have been received, carried out, or caused to be carried out, by the same server (i.e., the same software and / or hardware); it is intended to mean that any number of software elements or hardware devices may be involved in receiving / sending, carrying out or causing to be carried out any task or request, or the consequences of any task or request; and all of this software and hardware may be one server or multiple servers, both of which are included within the expressions “at least one server” and “a server”.

[0029] In the context of the present specification, “computing device” is any computing apparatus or computer hardware that is capable of running software appropriate to the relevant task at hand. Thus, some (non-limiting) examples of electronic devices include general purpose personal computers (desktops, laptops, notebooks, etc.), mobile computing devices, smartphones, and tablets, and network equipment such as routers, switches, and gateways. It should be noted that an electronic device in the present context is not precluded from acting as a server to other electronic devices. The use of the expression “an electronic device” does not preclude multiple electronic devices being used in receiving / sending, carrying out or causing to be carried out any task or request, or the consequences of any task or request, or steps of any method described herein. In the context of the present specification, a “client device” refers to any of a range of end-user client electronic devices, associated with a user, such as personal computers, tablets, smartphones, and the like.

[0030] In the context of the present specification, unless expressly provided otherwise, a computer system may refer, but is not limited to, an “electronic device”, a “client device”, a “computing device”, an “operation system”, a “system”, a “computer-based system”, a “computer system”, a “network system”, a “network device”, a “controller unit”, a “monitoring device”, a “control device”, a “server”, and / or any combination thereof appropriate to the relevant task at hand.

[0031] In the context of the present specification, the expression “computer readable storage medium” (also referred to as “storage medium” and “storage”) is intended to include non-transitory media of any nature and kind whatsoever, including without limitation RAM, ROM, disks (CD-ROMs, DVDs, floppy disks, hard drivers, etc.), USBkeys, solid state-drives, tape drives, etc. A plurality of components may be combined to form the computer information storage media, including two or more media components of a same type and / or two or more media components of different types.

[0032] In the context of the present specification, a “database” is any structured collection of data, irrespective of its particular structure, the database management software, or the computer hardware on which the data is stored, implemented or otherwise rendered available for use. A database may reside on the same hardware as the process that stores or makes use of the information stored in the database or it may reside on separate hardware, such as a dedicated server or plurality of servers. In the context of the present specification, the expression “information” includes information of any nature or kind whatsoever capable of being stored in a database. Thus, information includes, but is not limited to audiovisual works (images, movies, sound records, presentations etc.), data (location data, numerical data, etc.), text (opinions, comments, questions, messages, etc.), documents, spreadsheets, lists of words, etc.

[0033] In the context of the present specification, unless expressly provided otherwise, an “indication” of an information element may be the information element itself or a pointer, reference, link, or other indirect mechanism enabling the recipient of the indication to locate a network, memory, database, or other computer-readable medium location from which the information element may be retrieved. For example, an indication of a document could include the document itself (i.e., its contents), or it could be a unique document descriptor identifying a file with respect to a particular file system, or some other means of directing the recipient of the indication to a network location, memory address, database table, or other location where the file may be accessed. As one skilled in the art would recognize, the degree of precision required in such an indication depends on the extent of any prior understanding about the interpretation to be given to information being exchanged as between the sender and the recipient of the indication. For example, if it is understood prior to a communication between a sender and a recipient that an indication of an information element will take the form of a database key for an entry in a particular table of a predetermined database containing the information element, then the sending of the database key is all that is required to effectively convey the information element to the recipient, even though theinformation element itself was not transmitted as between the sender and the recipient of the indication.

[0034] In the context of the present specification, the expression “communication network” is intended to include a telecommunications network such as a computer network, the Internet, a telephone network, a Telex network, a TCP / IP data network (e.g., a WAN network, a LAN network, etc.), and the like. The term “communication network” includes a wired network or direct-wired connection, and wireless media such as acoustic, radio frequency (RF), infrared and other wireless media, as well as combinations of any of the above.

[0035] In the context of the present specification, the words “first”, “second”, “third”, etc. have been used as adjectives only for the purpose of allowing for distinction between the nouns that they modify from one another, and not for the purpose of describing any particular relationship between those nouns. Thus, for example, it should be understood that, the use of the terms “first server” and “third server” is not intended to imply any particular order, type, chronology, hierarchy or ranking (for example) of / between the servers, nor is their use (by itself) intended to imply that any “second server” must necessarily exist in any given situation. Further, as is discussed herein in other contexts, reference to a “first” element and a “second” element does not preclude the two elements from being the same actual real-world element. Thus, for example, in some instances, a “first” server and a “second” server may be the same software and / or hardware, in other cases they may be different software and / or hardware.

[0036] Implementations of the present technology each have at least one of the above-mentioned objects and / or aspects, but do not necessarily have all of them. It should be understood that some aspects of the present technology that have resulted from attempting to attain the above-mentioned object may not satisfy this object and / or may satisfy other objects not specifically recited herein.

[0037] Additional and / or alternative features, aspects and advantages of implementations of the present technology will become apparent from the following description, the accompanying drawings and the appended claims.BRIEF DESCRIPTION OF THE DRAWINGS

[0038] For a better understanding of the present technology, as well as other aspects and further features thereof, reference is made to the following description which is to be used in conjunction with the accompanying drawings, where:

[0039] FIG. 1 illustrates a flowchart of a computer-implemented method for estimating an air data parameter for an aircraft in accordance with one or more nonlimiting implementations of the present technology.

[0040] FIG. 2a illustrates a schematic diagram of a system for estimating an air data parameter for an aircraft in accordance with one or more non-limiting implementations of the present technology.

[0041] FIG. 2b illustrates a schematic diagram of a system for estimating an AoS, an AoA, a CAS and a Mach for an aircraft in accordance with one or more non-limiting implementations of the present technology.

[0042] FIG. 3 illustrates a schematic diagram of a Mach selection logic in accordance in accordance with one or more non-limiting implementations of the present technology.

[0043] FIG. 4 illustrates a flowchart of a computer-implemented method fortraining a set of neural networks to estimate an air data parameter for an aircraft in accordance with one or more non-limiting implementations of the present technology.

[0044] FIG. 5 illustrates a schematic diagram illustrating a training of a set of neural networks.

[0045] FIG. 6 illustrates a schematic diagram of a system for estimating an air data parameter for an aircraft in accordance with one or more non-limiting implementations of the present technology.

[0046] FIG. 7A illustrates a graph showing a CAS and an estimated CAS generated using a prior art method.

[0047] FIG. 7B illustrates a graph showing a difference between the CAS and the estimated CAS shown in Fig. 7A.

[0048] FIG. 8A illustrates a graph showing a CAS and an estimated CAS generated using an implementation of the present technology.

[0049] FIG. 8B illustrates a graph showing a difference between the CAS and the estimated CAS shown in Fig. 8A.DETAILED DESCRIPTION

[0050] The examples and conditional language recited herein are principally intended to aid the reader in understanding the principles of the present technology and not to limit its scope to such specifically recited examples and conditions. It will be appreciated that those skilled in the art may devise various arrangements which, although not explicitly described or shown herein, nonetheless embody the principles of the present technology and are included within its spirit and scope.

[0051] Furthermore, as an aid to understanding, the following description may describe relatively simplified implementations of the present technology. As persons skilled in the art would understand, various implementations of the present technology may be of a greater complexity.

[0052] In some cases, what are believed to be helpful examples of modifications to the present technology may also be set forth. This is done merely as an aid to understanding, and, again, not to define the scope or set forth the bounds of the present technology. These modifications are not an exhaustive list, and a person skilled in the art may make other modifications while nonetheless remaining within the scope of the present technology. Further, where no examples of modifications have been set forth, it should not be interpreted that no modifications are possible and / or that what is described is the sole manner of implementing that element of the present technology.

[0053] Moreover, all statements herein reciting principles, aspects, and implementations of the present technology, as well as specific examples thereof, are intended to encompass both structural and functional equivalents thereof, whether theyare currently known or developed in the future. Thus, for example, it will be appreciated by those skilled in the art that any block diagrams herein represent conceptual views of illustrative circuitry embodying the principles of the present technology. Similarly, it will be appreciated that any flowcharts, flow diagrams, state transition diagrams, pseudo-code, and the like represent various processes which may be substantially represented in computer-readable media and so executed by a computer or processor, whether or not such computer or processor is explicitly shown.

[0054] The functions of the various elements shown in the figures, including any functional block labeled as a “central processing unit”, may be provided through the use of dedicated hardware as well as hardware capable of executing software in association with appropriate software. When provided by a processor, the functions may be provided by a single dedicated processor, by a single shared processor, or by a plurality of individual processors, some of which may be shared. In some non-limiting implementations of the present technology, the processor may be a general-purpose processor, such as a central processing unit (CPU) or a processor dedicated to a specific purpose, such as a graphics processing unit (GPU). Moreover, explicit use of the term “processor” or “controller” should not be construed to refer exclusively to hardware capable of executing software, and may implicitly include, without limitation, digital signal processor (DSP) hardware, network processor, application specific integrated circuit (ASIC), field programmable gate array (FPGA), read-only memory (ROM) for storing software, random access memory (RAM), and non-volatile storage. Other hardware, conventional and / or custom, may also be included.

[0055] Software modules, or simply modules which are implied to be software, may be represented herein as any combination of flowchart elements or other elements indicating performance of process steps and / or textual description. Such modules may be executed by hardware that is expressly or implicitly shown.

[0056] The description thereof that follows is intended to be only a description of illustrative examples of the present technology. This description is not intended to define the scope or set forth the bounds of the present technology. In some cases, what are believed to be helpful examples of modifications to the environment 100 may alsobe set forth below. This is done merely as an aid to understanding, and, again, not to define the scope or set forth the bounds of the present technology. These modifications are not an exhaustive list, and, as a person skilled in the art would understand, other modifications are likely possible. Further, where this has not been done (i.e., where no examples of modifications have been set forth), it should not be interpreted that no modifications are possible and / or that what is described is the sole manner of implementing that element of the present technology. As a person skilled in the art would understand, this is likely not the case. In addition, it is to be understood that the environment 100 may provide in certain instances simple implementations of the present technology, and that where such is the case they have been presented in this manner as an aid to understanding. As persons skilled in the art would understand, various implementations of the present technology may be of a greater complexity.

[0057] With these fundamentals in place, some non-limiting implementations of the present technology will be considered.

[0058] The present technology relates to a method and system for estimating an air data parameter for an aircraft.

[0059] Method Description

[0060] Now referring to Fig. 1, there is shown a method 100 for estimating an air data parameter for an aircraft in accordance with one or more non-limiting implementations of the present technology. It will be appreciated that the method is implemented by a computer as further explained below. Accordingly, it is therefore also referred to as a computer-implemented method.

[0061] It will be appreciated that the air data parameter is selected from a group consisting of a true airspeed (TAS), a calibrated airspeed (CAS), a Mach number, an angle of attack (AoA) and a slide slip angle (AoS).

[0062] The method begins at processing step 102.

[0063] According to processing step 102, an estimation of a Mach regime for the aircraft is obtained by at least one processor.

[0064] It will be appreciated that various types of Mach regimes may be used. In fact it will be appreciated that according to the Mach, the flight envelope of the aircraft may be divided into various regimes known to the skilled addressee.

[0065] For instance, in one or more implementations, the Mach regime is one of a stall model, a subsonic model and a transonic model.

[0066] In one or more alternative implementations, the Mach regime is one of a stall model, a subsonic model, a transonic model and a supersonic model.

[0067] In one or more alternative implementations, the Mach regime comprises a plurality of predefined regimes. In one or more implementations, the predefined Mach regimes are characterized by intervals of 0.1 Mach. It will be appreciated by the skilled addressee that the number of divisions may be increased or decreased depending on the performance of the aircraft.

[0068] The skilled addressee will appreciate that various alternative implementations may be provided for the Mach regime.

[0069] It will be appreciated that the estimation of a Mach regime may be obtained according to various implementations. In one or more implementations, the estimation of a Mach regime is obtained via, for instance, the anemometer of the aircraft if the corresponding signal is valid. If the signal is not valid, the estimation of a Mach regime will be obtained from the neural network (NN).

[0070] According to processing step 104, an indication of at least one secondary surface configuration of the aircraft is obtained by the at least one processor.

[0071] The skilled addressee will also appreciate that the secondary surface depends on an aircraft model. For instance, for a given aircraft, the secondary surfaces may comprise a slat and a flap.

[0072] Moreover, it will be also appreciated by the skilled addressee that the configuration of the secondary surfaces also depends on the aircraft model. For instance, a given aircraft may be provided with three different configurations for the flaps while another model of aircraft may be provided with a different number ofpossible configurations for the flaps. In the case of the flaps, the configuration may be defined by a degree of extension.[00731 It will be appreciated that the indication of at least one secondary surface configuration of the aircraft may be obtained according to various implementations. In one or more implementations, the indication of at least one secondary surface configuration of the aircraft is obtained using a position sensor which provides an indication of corresponding position of the secondary surface. In one or more alternative implementations, the indication of at least one secondary surface configuration of the aircraft is obtained from a memory unit storing such indication.

[0074] According to processing step 106, a trained neural network is selected by the at least one processor. It will be appreciated that the trained model is selected based on the obtained estimation of the Mach regime and the indication of at least one secondary surface configuration. In fact, each trained neural network has been trained for a specific flight envelope so as to become an expert in performing prediction of an air data parameter for that specific flight envelope. Thus, processing step 106 aims to select the appropriate neural network to perform predictions for a given flight envelope based on the obtained estimation of the Mach regime and the indication of at least one secondary surface configuration. It will be appreciated that, in one or more implementations, each trained neural network has the same architecture and number of model parameters (e.g. weights) such that model parameters may be quickly retrieved and used for inference in the given flight envelope, which minimizes the computational load on the processing device executing the neural network.

[0075] In one or more implementations, the model parameters of the selected neural network are obtained from a non-transitory storage medium (e.g., a database). In one or more other implementations, the model parameters are obtained from another computing device operatively connected to the at least one processor.

[0076] Now referring to Fig. 2a, there is shown how the trained neural network is selected in accordance with one or more implementations.

[0077] A first switch 202 receives the obtained estimation of the Mach regime while a second switch 206 receives the indication of at least one secondary surface configuration, in this implementation, an indication of the flap and the slap configuration.

[0078] Depending on the various values, the first switch 202 and the second switch 206 select a set of weights 204 suitable to be used for the neural network 208.

[0079] It will be appreciated that in one or more implementations, the first switch 202 and the second switch 204 have a distinct 'on' and 'off position, in order to minimize flickering.

[0080] Now referring to Fig. 3, there is shown an implementation illustrating a logic selection for the first switch 202 receiving the estimation of the Mach regime in the implementation where the Mach regime is defined by 3 different Mach regime (ml, m2 and m3) and m2tol < mlto2 < m3to2< m2to3.

[0081] In this implementation, a first set reset gate 308 and a second set reset gate 310 are used. An additional block 312 is also used.

[0082] It will be appreciated by the skilled addressee that such configuration enables avoidance of flickering which is of great advantage. The skilled addressee will appreciate that various alternative implementations may be possible.

[0083] Now referring to Fig. 2b, there is shown a system for estimating an AoS, an AoA, a CAS and a Mach for an aircraft in accordance with one or more non-limiting implementations of the present technology.

[0084] In this implementation, the first switch 202 receives a Mach signal. The Mach signal received is originating from a Mach sensor if a signal indicating that the Mach signal received is valid and an estimated Mach signal received from the corresponding neural network if no signal indicating that the Mach signal received is valid.

[0085] The first switch 202 provides an indication of the Mach regime to each of a weight selection for AoS estimation unit, a weight selection for AoA estimation unit and a weight selection for airspeed estimation unit.

[0086] The weight selection for AoS estimation unit also receives from the second switch 206 an indication of at least one secondary surface configuration, in this implementation, an indication of the flap and the slap configuration and selects accordingly a set of weights for the neural network used for determining the AoS estimation.

[0087] The weight selection for AoA estimation unit also receives from the second switch 206 an indication of at least one secondary surface configuration, in this implementation, an indication of the flap and the slap configuration and selects accordingly a set of weights for the neural network used for determining the AoA estimation.

[0088] The weight selection for airspeed estimation unit also receives from the second switch 206 an indication of at least one secondary surface configuration, in this implementation, an indication of the flap and the slap configuration and selects accordingly a set of weights for the neural network used for determining the airspeed estimation.

[0089] The neural network for AoS estimation configured with the corresponding set of weights selected receives additional data suitable for estimating the AoS estimation and provides an estimation of the AoS.

[0090] The neural network for AoA estimation configured with the corresponding set of weights selected receives additional data suitable for estimating the AoA and provides an estimation of the AoA.

[0091] The neural network for airspeed estimation configured with the corresponding set of weights selected receives additional data suitable for estimating the airspeed provides an estimation of the airspeed. It will be appreciated that in one or more implementations both an estimation of the CAS and an estimation of the Mach are provided.

[0092] Now referring back to Fig. 1 and according to processing step 108, additional data suitable for estimating the air data parameter for the aircraft is obtained by the at least one processor.

[0093] It will be appreciated that the additional data suitable for estimating the air data parameter for the aircraft may be of various types.

[0094] In one or more implementations, the obtaining of the additional data suitable for estimating the air data parameter for the aircraft comprises obtaining an indication of a position of a horizontal primary surface of the aircraft.

[0095] The obtaining of the additional data suitable for estimating the air data parameter for the aircraft further comprises obtaining an estimation of a load exerted on a respective actuator for actuating the horizontal primary surface of the aircraft using at least the indication of the position of the horizontal primary surface and a position of the respective actuator.

[0096] The skilled addressee will appreciate that alternatively, the additional data suitable for estimating the air data parameter for the aircraft may further comprises an indication of a pitch, roll and yaw angle and rate associated with the aircraft. It will be appreciated that such data may be provided by the Inertial Reference System (IRS) for instance.

[0097] It will be appreciated that the additional data suitable for estimating the air data parameter for the aircraft may be obtained according to various implementations.

[0098] According to processing step 110, the selected trained neural network is accessed by the at least one processor.

[0099] In one or more implementations, the values of the model parameters (e.g., weights) of the selected trained neural network in the database are injected or set into the neural network executed by the processing device (e.g., LRU).

[0100] It will be appreciated that the selected trained neural network is accessed with at least the obtained additional data suitable for estimating the air data parameter for the aircraft.

[0101] In one or more implementations, the selected trained neural network is accessed with an indication of the at least one secondary surface configuration, the estimation of a load exerted on the respective actuator for actuating the horizontal primary surface of the aircraft, and the indication of the position of the horizontal primary surface of the aircraft.

[0102] The skilled addressee will appreciate that various alternative implementations may be possible.

[0103] According to processing step 112, an estimation of the air data parameter is obtained by the at least one processor.

[0104] It will be appreciated that the estimation of the air data parameter is obtained using the selected trained neural network based on the Mach regime, the indication of the at least one secondary surface configuration and the additional data.

[0105] According to processing step 114, the estimation of the air data parameter is provided by the at least one processor.

[0106] It will be appreciated that the estimation of the air data parameter for the aircraft may be provided according to various implementations.

[0107] In one or more implementations, the estimation of the air data parameter for the aircraft is provided to the Primary Flight Display (PFD) of the aircraft. It will be appreciated that the estimation of the air data parameter may also or alternatively be provided to the Flight Control Computer (FCC) of the aircraft.

[0108] The method 100 ends.

[0109] Now referring to Fig. 4, there is disclosed a method 400 for training a set of neural networks to estimate an air data parameter for an aircraft in accordance with one or more non-limiting implementations of the present technology.

[0110] It will be appreciated that the method is executed by at least one processor and is therefore also referred to as a computer-implemented method.

[0111] According to processing step 402, for each neural network of a set of neural networks, a respective training data set is obtained.[01121 It will be appreciated that in one or more implementations, each neural network has a same architecture and model parameters (i.e. same number of weights, same number of layers, same activation logic).

[0113] More precisely and in one or more implementations, each neural network has the following architecture. It is comprised of 3 layers de 60 neurons. It will be appreciated that an Elastic Net regulation is used. An Adam optimization is further used. It will be also appreciated that a ReLU activation function is also used.

[0114] The skilled addressee will appreciate that various alternative implementations may be possible for the architecture of the neural network.

[0115] It will be appreciated that each respective training data set comprises a respective plurality of training examples. Each respective training dataset corresponds to training dataset for a different flight envelope for which a respective neural network will be trained. Thus, a given neural network will become an expert for a respective flight envelope.

[0116] Each training example comprises a respective Mach regime, a respective indication of at least one secondary surface configuration of the aircraft, respective additional data suitable for estimating the air data parameter for the aircraft, and a respective air data parameter for the aircraft.

[0117] According to processing step 404, each neural network of the set of neural networks is trained on the respective training data set. It will be appreciated that the purpose of the training is to determine the respective air data parameter based on the respective Mach regime, the respective indication of at least one secondary surface configuration of the aircraft, and the respective additional data suitable for estimating the air data parameter for the aircraft.

[0118] Each neural network may be trained separately on its respective training dataset associated with a given flight envelope. It will be appreciated by the skilledaddressee that the training of each neural network of the set of neural networks may be performed in parallel or in sequence at different moments in time. It will be appreciated that the training may be performed by one or more processing devices, such as, but not limited to, a CPU, GPU, NPU, TPU, etc.

[0119] According to processing step 406, a set of trained neural networks is outputted. It will be appreciated that each trained neural network is configured to estimate the air data parameter of the aircraft according to a respective Mach regime.

[0120] In one or more implementations, the respective model parameters of each neural network may be stored in a storage medium (e.g. database) together with an indication of the respective flight envelope for which it has been trained.

[0121] Now referring to Fig. 5, there is shown an embodiment illustrating a neural network training phase for n data sets.

[0122] It will be appreciated that the neural network training phase may be performed by a server, not shown.

[0123] The server obtains the initial neural network by initializing the model parameters and model hyperparameters thereof.

[0124] The model parameters are configuration variables of the model used to perform predictions and which are estimated or learned from training data, i.e. the coefficients are chosen during learning based on an optimization strategy for outputting a prediction. The hyperparameters are configuration variables of a model which determine the structure of the initial model and how the initial model is trained.

[0125] It will be appreciated that the number of model parameters to initialize will depend on inter alia the type and the architecture of the model and the model hyperparameters (e.g. a number of layers, type of layers, number of neurons in a NN).

[0126] In one or more implementations, the hyperparameters include one or more of: a number of hidden layers and units, an optimization algorithm, a learning rate, momentum, an activation function, a minibatch size, a number of epochs, and dropout.

[0127] In one or more implementations, the hyperparameters are provided to the server. In one or more alternative implementations, the hyperparameters are initialized using one or more of an arbitrarily predetermined search, a grid search, a random search and Bayesian optimization.

[0128] In one or more implementations, for a model having a neural network-based architecture, the server initializes a set of weights characterizing the initial model. It will be appreciated that the manner in which the model parameters are initialized is not limited. In one or more implementations, the server initializes the model parameters randomly.

[0129] It will be appreciated that the supervised learning procedure is configured to inter alia, (i) obtain the set of initial models; (ii) obtain, for each initial model of the set of initial models, a respective training dataset; and (iii) train each initial model of the set of initial models on the respective training dataset to obtain a respective trained model of the set of trained model models.

[0130] During training, each model generates, for the feature vector of the given labelled example in the respective training dataset, a respective prediction.

[0131] Each respective prediction is compared to the label of the given labelled example and a respective loss is determined. The respective loss is determined by using a loss function. It will be appreciated that the choice of loss function depends on the type of prediction task. In one or more other implementations, for regression tasks, a mean squared error (MSE) loss function may be used. It will be appreciated that other types of loss functions known in the art may be used.

[0132] In one or more implementations, the server uses gradient-descent techniques.

[0133] The respective model is updated based on the respective loss calculated using an objective function.

[0134] The supervised learning procedure is repeated for each ML model of the set of ML models on the respective labelled training dataset until convergence to obtain a respective trained ML model of the set of trained ML models. As a non-limitingexample, the supervised learning procedure may stop upon reaching one or more of: a desired performance threshold (e.g. accuracy for classification tasks), a computing budget, a maximum training duration, a lack of improvement in performance, a system failure, and the like.

[0135] In implementations where the initial model is implemented as a NN-based model, the server updates the initial weights of the initial model incrementally after each training iteration (e.g. pass over a minibatch in the first training dataset) to obtain first trained weights of the first trained model.

[0136] The supervised learning procedure outputs a set of trained ML models, where each trained ML model is configured to estimate an air data parameter for an aircraft based a respective flight envelope.

[0137] A first data set 502 is used in a training phase 504 and a set of corresponding weights 506 is generated accordingly.

[0138] A second data set 508 is used in a training phase 510 and a set of corresponding weights 512 is generated accordingly.

[0139] A nthdata set 514 is used in a training phase 516 and a set of corresponding weights 518 is generated accordingly.

[0140] It will be appreciated that training different neural networks with specific training data which depends on the geometry of the wing as well as the airspeed of the aircraft will make it possible to make each neural network expert in a limited flight envelop, which will increase its performance, i.e. outputs accuracies, while controlling its size which is of great advantage.

[0141] Now referring to Fig. 6, there is shown a system 600 for estimating an air data parameter for an aircraft in accordance with one or more non-limiting implementations of the present technology.

[0142] The system 600 comprises a communication port 602. The communication port 602 is used to transfer data to and from the system 600 and may include interfaces for input devices, network interfaces for connecting to a network for instance. Theskilled addressee will appreciate that many implementations are possible for the communication port 602. In one or more implementations, the communication port 602 is operating with the Aeronautical Radio INC (ARINC) 429 protocol. It will be appreciated that the communication port 602 may also operate with the ARINC 664, the RS485 protocol or any manufacturer dependent protocol.

[0143] The system 600 further comprises a central processing unit 604.

[0144] It will be appreciated that the central processing unit 604 may be of various types. In one or more implementations, the central processing unit 604 used enables a real time processing and all computations are performed during the computation base cycle of the central processing unit 604, which is either 5ms or 10ms in one or more implementations.

[0145] The system 600 further comprises a memory unit 606. It will be appreciated that the memory unit 606 may be of various types.

[0146] More generally, the skilled addressee will appreciate that there are various types of memory unit 606, including random access memory (RAM), read-only memory (ROM) and flash memory. Alternative memory technologies include magnetic disk storage, optical disk storage, and solid-state storage.

[0147] It will be appreciated that the memory unit 606 comprises an application for estimating an air data parameter for an aircraft.

[0148] The application for estimating an air data parameter for an aircraft comprises instructions for obtaining an estimation of a Mach regime.

[0149] The application for estimating an air data parameter for an aircraft further comprises instructions for obtaining an indication of at least one secondary surface configuration.

[0150] The application for estimating an air data parameter for an aircraft further comprises instructions for selecting a trained neural network based on the obtained estimation of a Mach regime and the indication of at least one secondary surface configuration.

[0151] The application for estimating an air data parameter for an aircraft further comprises instructions for obtaining additional data suitable for estimating the air data parameter for the aircraft.

[0152] The application for estimating an air data parameter for an aircraft further comprises instructions for accessing the selected trained neural network with at least the obtained additional data suitable for estimating the air data parameter for the aircraft.

[0153] The application for estimating an air data parameter for an aircraft further comprises instructions for obtaining an estimation of the air data parameter.

[0154] The application for estimating an air data parameter for an aircraft further comprises instructions for providing the estimation of the air data parameter.

[0155] There is also disclosed a system for estimating an air data parameter for an aircraft.

[0156] The system comprises one or more processing devices configured for obtaining an estimation of a Mach regime.

[0157] The one or more processing devices are further configured for obtaining an indication of at least one secondary surface configuration.

[0158] The one or more processing devices are further configured for selecting a trained neural network based on the obtained estimation of a Mach regime and the indication of at least one secondary surface configuration.

[0159] The one or more processing devices are further configured for obtaining additional data suitable for estimating the air data parameter for the aircraft.

[0160] The one or more processing devices are further configured for accessing the selected trained neural network with at least the obtained additional data suitable for estimating the air data parameter for the aircraft.

[0161] The one or more processing devices are further configured for obtaining an estimation of the air data parameter.[01621 The one or more processing devices are further configured for providing the estimation of the air data parameter.

[0163] It will be appreciated that in one or more implementations, the method for estimating an air data parameter for an aircraft in accordance with one or more nonlimiting implementations of the present technology is integrated into the LRU of the aircraft.

[0164] Now referring to Fig. 7A, there is shown superimposed for a flight a CAS and a CAS estimated according to the prior art.

[0165] Fig, 7B shows for that flight disclosed in Fig. 7A a difference over time between the CAS and the CAS estimated according to the prior art. The difference is indicative of an error between the CAS and the CAS estimated according to the prior art.

[0166] Fig. 8A shows for that same flight the CAS and a CAS estimated according to an implementation of the technology disclosed herein.

[0167] Fig. 8B shows for that same flight a difference over time between the CAS and the CAS estimated according to an implementation of the technology disclosed herein. The difference is indicative of an error between the CAS and the CAS estimated according to an implementation of the technology disclosed herein.

[0168] The skilled addressee will readily appreciate the advantages of the technology disclosed herein by comparing the error illustrated in Fig. 8B to the error illustrated in Fig. 7B.

[0169] Moreover, the skilled addressee will appreciate that the technology disclosed herein enables an integration in the LRU despites the limitations associated with the LRU which is of great advantage.

[0170] It will be appreciated that there is also disclosed a non-transitory machine-readable medium carrying instructions, which, when executed by at least one processor, perform one or more implementations of the computer-implemented methods disclosed above.

[0171] In some cases, what are believed to be helpful examples of modifications to the present technology may also be set forth. This is done merely as an aid to understanding, and, again, not to define the scope or set forth the bounds of the present technology. These modifications are not an exhaustive list, and a person skilled in the art may make other modifications while nonetheless remaining within the scope of the present technology. Further, where no examples of modifications have been set forth, it should not be interpreted that no modifications are possible and / or that what is described is the sole manner of implementing that element of the present technology.

[0172] Modifications and improvements to the above-described implementations of the present technology may become apparent to those skilled in the art. The foregoing description is intended to be exemplary rather than limiting.

Claims

CLAIMS:

1. A computer-implemented method for estimating an air data parameter for an aircraft, the method being executed by at least one processor, the method comprising:obtaining an estimation of a Mach regime for the aircraft; obtaining an indication of at least one secondary surface configuration of the aircraft;selecting a trained neural network based on the obtained estimation of the Mach regime and the indication of at least one secondary surface configuration;obtaining additional data suitable for estimating the air data parameter for the aircraft;accessing the selected trained neural network with at least the obtained additional data suitable for estimating the air data parameter for the aircraft;obtaining, using the selected trained neural network based on the Mach regime, the indication of the at least one secondary surface configuration and the additional data, an estimation of the air data parameter; andproviding the estimation of the air data parameter.

2. The computer-implemented method as claimed in claim 1, wherein the obtaining of the additional data suitable for estimating the air data parameter for the aircraft comprises:obtaining an indication of a position of a horizontal primary surface of the aircraft;obtaining an estimation of a load exerted on a respective actuator for actuating the horizontal primary surface of the aircraft using at least the indication of the position of the horizontal primary surface and a position of the respective actuator;wherein the selected trained neural network is accessed with an indication of the at least one secondary surface configuration, the estimation of a load exerted on the respective actuator for actuating the horizontal primary surface of the aircraft, and the indication of the position of the horizontal primary surface of the aircraft.

3. The computer-implemented method as claimed in any one of claims 1 to 2, wherein the air data parameter is selected from a group consisting of a true airspeed (TAS), a calibrated airspeed (CAS), a Mach number, an angle of attack (AoA) and a slide slip angle (AoS).

4. The computer-implemented method as claimed in any one of claims 1 to 3, wherein the Mach regime is one of a stall model, a subsonic model, a transonic model and a supersonic.

5. The computer-implemented method as claimed in any one of claims 1 to 4, wherein said selecting the trained neural network based on the obtained estimation of the Mach regime and the indication of at least one secondary surface configuration comprises:selecting the trained neural network from a set of trained neural networks, each trained neural network being configured to estimate the air data parameter according to at least a respective Mach regime, a respective range of secondary surface configurations and a respective range of additional data suitable for estimating the air data parameter for the aircraft.

6. The computer-implemented method as claimed in claim 5, wherein each trained neural network has a same architecture.

7. The computer-implemented method as claimed in any one of claims 1 to 6, wherein the at least one processor is part of a line remote unit (LRU).

8. A computer-implemented method for training a set of neural networks to estimate an air data parameter for an aircraft, the method being executed by at least one processor, the method comprising:obtaining, for each neural network of the set of neural networks, a respective training data set comprising a plurality of training examples, each training example comprising:a respective Mach regime,a respective indication of at least one secondary surface configuration of the aircraft,respective additional data suitable for estimating the air data parameter for the aircraft, anda respective air data parameter for the aircraft;training each neural network of the set of neural networks on the respective training data set to determine the respective air data parameter based on the respective Mach regime, the respective indication of at least one secondary surface configuration of the aircraft, and the respective additional data suitable for estimating the air data parameter for the aircraft; and outputting a set of trained neural networks, each trained neural network being configured to estimate the air data parameter of the aircraft according to a respective Mach Regime.

9. The computer-implemented method as claimed in claim 5, wherein each trained neural network of the set of trained neural networks has a same architecture.

10. A system for estimating an air data parameter for an aircraft, the system comprising:one or more processing devices configured for:obtaining an estimation of a Mach regime;obtaining an indication of at least one secondary surface configuration;selecting a trained neural network based on the obtained estimation of a Mach regime and the indication of at least one secondary surface configuration;obtaining additional data suitable for estimating the air data parameter for the aircraft;accessing the selected trained neural network with at least the obtained additional data suitable for estimating the air data parameter for the aircraft;obtaining an estimation of the air data parameter; and providing the estimation of the air data parameter.

11. A processing device comprising:a central processing unit;a communication port;a memory unit comprising an application for estimating an air data parameter for an aircraft, the application comprising instructions for:obtaining an estimation of a Mach regime;obtaining an indication of at least one secondary surface configuration; selecting a trained neural network based on the obtained estimation of a Mach regime and the indication of at least one secondary surface configuration;obtaining additional data suitable for estimating the air data parameter for the aircraft;accessing the selected trained neural network with at least the obtained additional data suitable for estimating the air data parameter for the aircraft;obtaining an estimation of the air data parameter; andproviding the estimation of the air data parameter.

12. A non-transitory machine-readable medium carrying instructions, which, when executed by at least one processor, perform the computer-implemented method as claimed in any one of claims 1 to 9.

Images