System Integration

Abstract

A computer-implemented method for generating, while in flight, a feasibility indication indicating the feasibility of a weapon carried on the aircraft successfully engaging a target and / or the feasibility of a weapon carried on the target successfully engaging the aircraft, the method comprising the steps of providing a database describing a performance envelope of the weapon, identifying a best candidate polynomial from a plurality of candidate polynomials, using a generic algorithm to generate coefficients characteristic of that performance envelope, wherein the generic algorithm has a polynomial form, the polynomial variables being some or all of a group of firing condition parameters of the weapon or the aircraft, and transmitting the coefficients of the identified best candidate polynomial to the aircraft. 1. A method comprising: uploading a generic algorithm to a reconfigurator that includes a same generic algorithm; selecting, by an on-aircraft reconfigurator that includes the same generic algorithm, coefficients for the generic algorithm according to aircraft and target conditions; and generating, by the reconfigurator using the selected coefficients, a feasibility indication, wherein selecting, by an on-aircraft reconfigurator that includes the same generic algorithm, coefficients for the generic algorithm according to aircraft and target conditions comprises selecting, by an on-aircraft reconfigurator that includes the same generic algorithm, coefficients for the generic algorithm according to aircraft and target conditions if the aircraft and target are within the weapon's performance envelope.

Description

[Technical Field]

[0001] This invention relates to system integration, and more specifically, to the integration of complex and highly integrated weapons systems on aircraft. [Background technology]

[0002] Integrating weapon systems with other systems on an aircraft is a complex and redundant task because it affects all major aircraft systems. Therefore, there is a need to improve the time and affordability of weapon integration.

[0003] One of the things required for weapons integration is that weapons can target specific targets and Successfully engage in combat The purpose is to enable aircraft pilots to see information regarding whether or not it is possible. For this purpose, weapons are usually grouped into two categories: weapons designed to engage ground targets (air-to-ground weapons) and weapons designed to engage airborne targets (air-to-air weapons). In the case of air-to-ground weapons, the selected target and Successfully engage in combat Alternatively, a Launch Acceptability Region (LAR) is calculated, which is the region where the probability of hitting the target exceeds a certain threshold. The LAR is calculated based on the target and Successfully engage in combat Calculated to provide a cockpit display indicating feasibility during launch, it is a function of weapon performance characteristics, the relative position and motion of the aircraft and the target, and often ambient conditions such as wind speed and direction.

[0004] In the case of air-to-air weapons, the selected aerial target and Successfully engage in combat A Launch Success Zone (LSZ) is calculated, indicating that the probability of success exceeds a certain threshold. In this case as well, the LSZ is calculated when the weapon has met its target. Successfully engage in combat It is used to provide a cockpit display indicating whether it is possible. However, calculating the LSZ is more complex than calculating the LAR because the relative speed and direction of the launching aircraft and the target are far more significant, the influence of surrounding conditions is greater, and the physical properties of the weapon in flight are more important in the calculation.

[0005] The conventional approach involved creating a simple abstract model of the weapon, modified according to launch conditions (taking into account aircraft and target conditions (e.g., range, direction, and speed of travel, etc.) and surrounding conditions). The model was then mounted on the aircraft and used to generate a LAR or LSZ for display to the pilot. The drawback of the conventional approach is that each model is different for each different weapon type. Storing data on several different implicit models consumes considerable memory capacity, and each model must be comprehensively integrated to ensure that it does not negatively impact any of the aircraft systems. Furthermore, if any changes or modifications are made to the weapon (such as performance improvements), or if it is necessary to load an entirely new weapon onto the aircraft, a redundant and costly integration process must be undertaken because the weapon model will be substantially different from any previously integrated with the aircraft system. [Overview of the project]

[0006] According to a first aspect of the present invention, in an aircraft in flight, a weapon carried on the aircraft targets a target Successfully engage in combat The feasibility and / or the weapons that can be delivered on the target are with aircraft and Successfully engage in combat A computer implementation method is provided for generating a feasibility statement indicating feasibility, and this method is To provide a database that describes the performance envelope of weapons, The process involves creating the coefficient characteristics of its performance envelope using a general algorithm, which includes a step of identifying the best candidate polynomial from several candidate polynomials, wherein the general algorithm has the form of a polynomial, and the variables of the polynomial are some or all of a group of firing condition parameters for a weapon or aircraft. Uploading the coefficients of the best identified candidate polynomial to the aircraft, The reconfigurator on the aircraft, which includes the same general algorithm, selects coefficients for the general algorithm according to the conditions of the aircraft and the target, Using the selected coefficients, the reconstructor generates a feasibility representation. A method comprising the same general algorithm, wherein the reconfigurator on the aircraft, which includes the same general algorithm, selects coefficients for the general algorithm according to the conditions of the aircraft and the target, provided that the aircraft and the target are within the performance envelope of the weapon.

[0007] In this way, a capability filter (CF) is provided to find the limits of the possible envelopes of weapon systems within an arbitrarily selected region of the engagement envelope and to examine the feasibility of weapon engagement under current firing conditions, i.e., to determine whether the current firing is inside or outside the "zone of impact" for the weapon. This provides the prospect of classifying a large, high-dimensional space using a relatively simple model, thereby saving both processing and memory for the host system. Firstly, the capability filter assesses whether the weapon is capable or not. Secondly, if the weapon is capable, the relevant LSZ / LAR parameters are estimated.

[0008] In one example, the method involves using a trained machine learning model, such as a trained neural network, to infer whether an aircraft and a target are within the performance envelope of a weapon, according to the conditions of the aircraft and the target.

[0009] In this way, it is possible to determine whether an aircraft and a target are within the performance envelope of a weapon, according to the conditions of the aircraft and the target.

[0010] In one example, this method involves training a machine learning model using training data of the performance envelopes of each weapon, according to the conditions of each aircraft and each target.

[0011] In this way, the machine learning model can be trained using the training data of the performance envelope of each weapon according to each aircraft and each target condition.

[0012] In one example, the method comprises labeling the training data based on whether each aircraft and each target are within the performance envelope of each weapon according to each aircraft and each target condition.

[0013] In this way, the training data can be labeled, for example, capable = 1, incapable = 0.

[0014] In one example, the method comprises creating the coefficient characteristics of each performance envelope using a general algorithm by steps including identifying the best candidate polynomial from a plurality of candidate polynomials, where the variables of the polynomial are some or all of a group of firing condition parameters of each weapon or aircraft.

[0015] In this way, each coefficient is created for the training data, for example, in the same way as the performance envelope of the weapon.

[0016] In one example, inferring whether an aircraft and a target are within the performance envelope of a weapon using a trained machine learning model according to aircraft and target conditions comprises thresholding the result of the inference.

[0017] In this way, a binary output, for example, capable = 1, incapable = 0, can be determined from the result of the inference.

[0018] In one example, selecting the coefficients for a general algorithm according to aircraft and target conditions when the aircraft and the target are within the performance envelope of the weapon by a reconstructor on the aircraft including the same general algorithm comprises selecting the coefficients for the general algorithm by a reconstructor on the aircraft including the same general algorithm when the aircraft and the target are currently within the performance envelope of the weapon.

[0019] In this way, when the aircraft and the target are within the performance envelope of the weapon, coefficients for the general algorithm are selected, thereby optimizing the coefficients for the general algorithm, and thus, according to the conditions of the aircraft and the target, the weapon carried on the aircraft and the target Successfully engage in combat feasibility and / or the weapon carried on the target and the aircraft Successfully engage in combat improve the determination of feasibility.

[0020] In one example, selecting coefficients for the general algorithm according to the conditions of the aircraft and the target when the aircraft and the target are within the performance envelope of the weapon by a reconfigurer on the aircraft that includes the same general algorithm comprises selecting coefficients for the general algorithm only when the aircraft and the target are within the performance envelope of the weapon by a reconfigurer on the aircraft that includes the same general algorithm.

[0021] In this way, coefficients for the general algorithm are selected only when the aircraft and the target are within the performance envelope of the weapon, thereby optimizing the coefficients for the general algorithm, and thus, according to the conditions of the aircraft and the target, the weapon carried on the aircraft and the target Successfully engage in combat feasibility and / or the weapon carried on the target and the aircraft Successfully engage in combat improve the determination of feasibility.

[0022] In one example, selecting coefficients for the general algorithm according to the conditions of the aircraft and the target when the aircraft and the target are within the performance envelope of the weapon by a reconfigurer on the aircraft that includes the same general algorithm comprises selecting coefficients for the general algorithm while the aircraft and the target are within the performance envelope of the weapon by a reconfigurer on the aircraft that includes the same general algorithm.

[0023] In this way, while the aircraft and target are within the performance envelope of the weapon, coefficients for the general algorithm are selected, thereby optimizing the coefficients for the general algorithm, and thus, according to the conditions of the aircraft and target, the weapon carried on the aircraft is able to target the target. Successfully engage in combat The feasibility and / or the weapons that can be delivered on the target are with aircraft and Successfully engage in combat Improve feasibility determination.

[0024] For example, selecting coefficients for a general algorithm according to the conditions of the aircraft and target, when the aircraft and target are within the performance envelope of the weapon, by a reconfigurator on an aircraft containing the same general algorithm, comprises repeatedly selecting coefficients for a general algorithm when the aircraft and target are within the performance envelope of the weapon, by a reconfigurator on an aircraft containing the same general algorithm.

[0025] In this way, when the aircraft and target are within the performance envelope of the weapon, coefficients for the general algorithm are repeatedly selected, for example periodically (e.g., on an ms timescale) or intermittently, thereby repeatedly optimizing the coefficients for the general algorithm, and thus, according to the conditions of the aircraft and target, the weapon carried on the aircraft and the target are Successfully engage in combat The feasibility and / or the weapons that can be delivered on the target are with aircraft and Successfully engage in combat Improve feasibility determination.

[0026] For example, a reconfigurator on an aircraft containing the same general algorithm selects coefficients for the general algorithm according to the conditions of the aircraft and target when the aircraft and target are within the performance envelope of the weapon, and a reconfigurator on an aircraft containing the same general algorithm deselects coefficients for the general algorithm when the aircraft and target are no longer within the performance envelope of the weapon.

[0027] In this way, if the aircraft and target are no longer within the performance envelope of the weapon, the feasibility indication is not generated by the reconfigurator.

[0028] One example involves creating the coefficient characteristics of its performance envelope using a general algorithm, which includes a step of identifying the best candidate polynomial from several candidate polynomials, where the general algorithm has the form of a polynomial, and the variables of the polynomial are some or all of a group of firing condition parameters for a weapon or aircraft. a) Generate candidate polynomials, where the variables of the polynomials are some or all of the group of firing condition parameters for weapons or aircraft. b) For each candidate polynomial, use the least-squares error criterion to calculate the coefficients for the candidate polynomial that best fits the characteristics of the weapon's performance envelope, c) For each candidate polynomial, a candidate score is generated according to the quality of the fit of that candidate polynomial to the characteristics of the weapon's performance envelope, d) Apply a genetic algorithm to the candidate polynomials and scores, including selecting the best-scoring polynomial(s) and discarding the other polynomial(s), thereby identifying the best candidate polynomial and its coefficients. e) Repeat the identification process until all required properties of the performance envelope have corresponding polynomial models. The process involves a step that includes creating the coefficient characteristics of its performance envelope using a general algorithm, which has the form of a polynomial.

[0029] For example, the types of candidate polynomials in a set of candidate polynomials include single-variable polynomials, multi-variable polynomials, and modified forms thereof. Other polynomial types are known.

[0030] For example, the order of the candidate polynomials in the set of candidate polynomials is in the range of 1 to 100, preferably in the range of 2 to 25, more preferably in the range of 3 to 10, and most preferably in the range of 5 to 9, for example, 5, 6, 7, 8, 9.

[0031] For example, a general polynomial takes the following form:

[0032]

number

[0033] Here: α mn This represents the m coefficients required to calculate the output n, {x1…x Ni} represents the normalized input, {y1…y Ni} represents the output, p 1mn This represents the power (exponent) of the m-th term of the n-th polynomial by the x1 variable.

[0034] In one example, the best candidate polynomial is of the following form:

[0035]

number

[0036] Here:

[0037]

number

[0038]

number

[0039]

number

[0040] In one example, the order of a general polynomial is 3 or greater. In another example, the order of a general polynomial is in the range of 10 to 25, for example, 20. Surprisingly, the inventors have found that using a general algorithm with an order of about 20 accurately and sufficiently describes most air-to-air engagements at a suitable runtime for an on-air implementation. Nevertheless, general algorithms can have an order greater than 2.

[0041] In one example, step b) calculates the coefficients for each candidate polynomial that best fits that candidate polynomial to the characteristics of the weapon's performance envelope, using the least-squares error criterion, and comprises: 1) generating an initial population of candidate polynomials; 2) for each candidate polynomial, calculating a set of coefficients that fits that polynomial to the performance envelope according to one or more criteria; 3) for each set of candidate polynomials and coefficients, calculating a score function indicating the quality of the fit of that set of candidate polynomials and coefficients to the performance envelope; and 4) recursively applying a genetic algorithm to the set of candidate polynomials until one or more criteria are met, including retaining at least the best-scoring polynomials and discarding the other polynomials. In one example, the output of the retained polynomials is a layer of a Self-Organizing Polynomial Neural Network, which is used to provide input for creating higher-order candidate polynomials. In one example, these steps are repeated for higher-order candidate polynomials. In another example, the final result is obtained from the path that ends with the best candidate score.

[0042] In this way, the method significantly improves weapon integration time and cost. More specifically, typically, the genetic algorithm proceeds iteratively by generating a new population of strings from a population of old strings. All strings are encoded versions of provisional solutions. An evaluation function associates a goodness-of-fit measure with each string, indicating its suitability for the problem. The algorithm applies probabilistic operators such as selection, crossover, and mutation to a random population to compute the entire generation of new strings. The inventors have identified that these algorithms can be adapted for use on multiple processor workstations or distributed systems with transparent process transitions. All goodness-of-fit evaluations and fitting operations can be performed in separate processes, i.e., simultaneously. For trivial goodness-of-fit functions, due to the level of overhead, little improvement in the rate of evolution is likely to be observed. However, for weapon targeting and many tasks on a particular system, this is clearly time-consuming and reflects the ad-hoc nature of the solution process. The accuracy of the goodness-of-fit depends on the complexity of the performance envelope. To accurately represent the performance envelope and achieve the required accuracy for the entire envelope or a subset thereof, a large dataset may be necessary. As the data size increases, so do the computational complexity and processing time. The benefits of performing goodness-of-fit evaluations in parallel are significant.

[0043] To apply a genetic algorithm to a specific purpose, an internal representation of the space to be explored is selected, and an external function is defined that assigns goodness-of-fit values ​​to candidate solutions.

[0044] This method can be used for different weapon types, and each set of coefficients can be easily determined for each weapon type, for example, for each of several different firing conditions (i.e., aircraft and target conditions). Aircraft and target conditions may include, but are not limited to, one or more of their relative position, distance, direction of movement, speed, and ambient atmospheric conditions. Weapon or aircraft firing condition parameters may include, but are not limited to, parameters such as aircraft speed, aircraft altitude, aircraft attitude, slant range to target, target speed, target altitude, azimuth angle of line of sight, target pitch and aspect angle, and wind speed. Weapon or aircraft firing condition parameters may include, but are not limited to, the relative speed and direction of the launching aircraft and target, and the relative speed and direction of the weapon's progress relative to the target.

[0045] Advantageously, the general polynomials / algorithms described above can be used by multiple different types of aircraft (e.g., simultaneously). In other words, different types of aircraft can use the same general algorithm to calculate LAR / LSZ. Also, the same general algorithm can be used to calculate LAR / LSZ for different weapon types. This means that aircraft software with general polynomials and means to enable loading coefficients for each weapon loaded on the aircraft is created only once. The software algorithms and coefficients for any given weapon are the same for any aircraft type. This is because common tools can be used for polynomial and coefficient generation, but both the software (including algorithms / polynomials) and coefficients tend to be generated for each weapon type and each time weapon performance changes, unlike the conventional method. This need to rewrite the software and its certification tends to be particularly costly. The methods and systems described above tend to offer the advantage that the aircraft software does not need to be rewritten and therefore does not require new certification.

[0046] In some embodiments, each aircraft in a fleet consisting of multiple different aircraft is loaded with the same common general polynomial. When a weapon is loaded onto an aircraft in the fleet, specific coefficients corresponding to that weapon may also be loaded onto that aircraft. This is in contrast to conventional systems where the tools for generating LAR / LSZ may be common across multiple different aircraft, but when a weapon is loaded onto an aircraft, both the polynomial / algorithm for generating LAR / LSZ and the corresponding coefficients are generated for the aircraft and the loading and unloading of the weapon.

[0047] The coefficients can be implemented as loadable data to enable accurate and precise weapon behavior within the weapon system. Furthermore, using only one or a few common algorithms would allow different weapon systems to be authorized or certified / approved with reduced effort and more quickly than the extensive testing required by conventional approaches, for use with aircraft. In other words, a minimum number of common weapon targeting algorithms may be used to account for all weapon types.

[0048] The use of common algorithms for weapon targeting also allows for increased or significant changes in weapon system capabilities to be integrated with aircraft systems with significantly less effort than before.

[0049] Weapons being transported on the target are aircraft and Successfully engage in combat By determining feasibility, the aircraft will be able to withstand the weapons carried by the hostile target. Successfully engage in combat It displays whether or not there is a risk, and to what extent. This calculation of the opposing LSZ / MEZ allows for a better assessment of the engagement. This, in turn, can lead to a more confident prediction of the advantages and potential outcomes of the engagement.

[0050] Advantageously, the above embodiments provide a general polynomial / algorithm that can be used by multiple different types of aircraft (e.g., simultaneously). Different types of aircraft may use the same general algorithm to calculate LAR / LSZ. Also, the same general algorithm may be used to calculate LAR / LSZ for different weapon types. Thus, aircraft software with a general polynomial and means to enable loading coefficients for each weapon loaded on the aircraft is created only once. The software algorithm and coefficients for any given weapon are the same for any aircraft type. This is because common tools may be used for polynomial and coefficient generation, but both the software (including algorithms / polynomials) and coefficients tend to be generated for each weapon type and each time weapon performance changes, unlike the conventional method. This need to rewrite the software and its certification tends to be particularly costly. The methods and systems described above tend to advantageously provide that the aircraft software does not need to be rewritten and therefore does not require new certification.

[0051] In one example, the target is equipped with and / or is an aircraft. In another example, the feasibility indication shows the successful launch zone for the aircraft and / or the target.

[0052] In one example, the target includes and / or is a ground-based target. In another example, the feasibility indication shows the aircraft's launch area and / or the target's missile engagement zone.

[0053] For example, step b) calculates the coefficients for each candidate polynomial that best fits the characteristics of the weapon's performance envelope, using the least-squares error criterion. The process involves generating an initial population of candidate polynomials, For each candidate polynomial, calculate a set of coefficients that fit the polynomial to the performance envelope according to one or more criteria (e.g., the least squares criterion), For each set of candidate polynomials and coefficients, calculate a score function that indicates the quality of the fit of that set of candidate polynomials and coefficients to the performance envelope. The genetic algorithm is applied recursively to the set of candidate polynomials until one or more criteria are met, including retaining at least the best-scoring polynomial and discarding other polynomials. It is equipped with.

[0054] In one example, the output of the retained polynomial(s) is a layer of a self-organizing polynomial neural network, used to provide input for constructing higher-order candidate polynomials. In one example, these are iterated over until the final result with the best candidate score is obtained.

[0055] In one example, the performance envelope of a weapon is the minimum envelope that defines the performance of the weapon when it is mounted on an aircraft, e.g., the minimum envelope that defines the performance of the weapon. In another example, the performance envelope of a weapon is the performance of the weapon when it is mounted on an aircraft of a different aircraft type. In another example, the method comprises obtaining the respective performance envelopes for one or more different aircraft types, e.g., multiple different aircraft types.

[0056] In one example, the method comprises determining a performance envelope using multiple aircraft performance envelopes, including determining a performance envelope that defines all the performance of different aircraft types (i.e., a "maximum aircraft performance envelope"), and using the performance envelopes and weapon performance envelopes representing all the performance of different aircraft types, determining a performance envelope that defines the performance of a weapon when it is implemented on each of the different aircraft types. In one example, the performance envelope is the smallest-sized envelope that defines the performance of a weapon when it is implemented on each of the different aircraft types.

[0057] In some embodiments, the database is generated by defining the range of conditions under which a weapon may be required to be fired, the range of aircraft conditions under which an aircraft can fire a weapon, and the range of weapon conditions under which a weapon can fire; generating data showing weapon performance for each weapon firing possibility from within the defined ranges; and creating a database that defines the overall performance envelope of the weapon. Coefficients can then be determined from this database and a general polynomial. In this way, the database can be generated on a ground-based system, and therefore the aircraft system only needs capacity to store a general polynomial and process the coefficients using aircraft and target conditions in order to generate the feasibility representation. As a result, the amount of data storage / processing capacity required on the aircraft tends to be reduced.

[0058] The coefficients can be implemented as loadable data to enable accurate and precise weapon behavior within the weapon system. Furthermore, using only one or a few common algorithms would allow different weapon systems to be authorized or certified / approved for use with aircraft with reduced effort and more quickly than the extensive testing required by conventional approaches.

[0059] The step of uploading the generated coefficients to the aircraft may be performed when the weapon is loaded as an aircraft containment. When a new weapon containment is loaded, the coefficients associated with that weapon may be uploaded to the aircraft at the same time as the weapon in order to integrate the weapon with the aircraft targeting system. Preferably, the coefficients are stored on a hardware device along with the weapon, and the device is connected to the aircraft to upload the coefficient data when the weapon is loaded.

[0060] For example, the types of candidate polynomials in a set of candidate polynomials include single-variable polynomials, multi-variable polynomials, and modified forms thereof. Other polynomial types are known.

[0061] For example, the order of the candidate polynomials in the set of candidate polynomials is in the range of 1 to 100, preferably in the range of 2 to 25, more preferably in the range of 3 to 10, and most preferably in the range of 5 to 9, for example, 5, 6, 7, 8, 9.

[0062] For example, a general polynomial takes the following form:

[0063]

number

[0064] Here: α mn This represents the m coefficients required to calculate the output n, {x1…x Ni} represents the normalized input, {y1…y Ni} represents the output, p 1mn This represents the power (exponent) of the m-th term of the n-th polynomial by the x1 variable.

[0065] In one example, the best candidate polynomial is of the following form:

[0066]

number

[0067] Here:

[0068]

number

[0069]

number

[0070]

number

[0071] In one example, the order of a general polynomial is 3 or greater. In another example, the order of a general polynomial is in the range of 10 to 25, for example, 20. Surprisingly, the inventors have found that using a general algorithm with an order of about 20 accurately and sufficiently describes most air-to-air engagements at a suitable runtime for an on-air implementation. Nevertheless, general algorithms can have an order greater than 2.

[0072] In one example, step b) calculates the coefficients for each candidate polynomial that best fits that candidate polynomial to the characteristics of the weapon's performance envelope, using the least-squares error criterion, and comprises: 1) generating an initial population of candidate polynomials; 2) for each candidate polynomial, calculating a set of coefficients that fits that polynomial to the performance envelope according to one or more criteria; 3) for each set of candidate polynomials and coefficients, calculating a score function indicating the quality of the fit of that set of candidate polynomials and coefficients to the performance envelope; and 4) recursively applying a genetic algorithm to the set of candidate polynomials until one or more criteria are met, including retaining at least the best-scoring polynomials and discarding the other polynomials. In one example, the output of the retained polynomials is a layer of a self-organizing polynomial neural network, which is used to provide input for creating higher-order candidate polynomials. In one example, these steps are iterated over the higher-order candidate polynomials. In one example, the final result is obtained from the path that ends with the best candidate score.

[0073] In one example, the target is equipped with and / or is an aircraft. In another example, the feasibility indication shows the successful launch zone for the aircraft and / or the target.

[0074] In one example, the target includes and / or is a ground-based target. In another example, the feasibility indication shows the aircraft's launch area and / or the target's missile engagement zone.

[0075] For example, step b) calculates the coefficients for each candidate polynomial that best fits the characteristics of the weapon's performance envelope, using the least-squares error criterion. The process involves generating an initial population of candidate polynomials, For each candidate polynomial, calculate a set of coefficients that fit the polynomial to the performance envelope according to one or more criteria (e.g., the least squares criterion), For each set of candidate polynomials and coefficients, calculate a score function that indicates the quality of the fit of that set of candidate polynomials and coefficients to the performance envelope. The genetic algorithm is applied recursively to the set of candidate polynomials until one or more criteria are met, including retaining at least the best-scoring polynomial and discarding other polynomials. It is equipped with.

[0076] In one example, the output of the retained polynomial(s) is a layer of a self-organizing polynomial neural network, used to provide input for constructing higher-order candidate polynomials. In one example, these are iterated over until the final result with the best candidate score is obtained.

[0077] In one example, the performance envelope of a weapon is the minimum envelope that defines the performance of the weapon when it is mounted on an aircraft, e.g., the minimum envelope that defines the performance of the weapon. In another example, the performance envelope of a weapon is the performance of the weapon when it is mounted on an aircraft of a different aircraft type. In another example, the method comprises obtaining the respective performance envelopes for one or more different aircraft types, e.g., multiple different aircraft types.

[0078] In one example, the method comprises determining a performance envelope using multiple aircraft performance envelopes, including determining a performance envelope that defines all the performance of different aircraft types (i.e., a "maximum aircraft performance envelope"), and using the performance envelopes and weapon performance envelopes representing all the performance of different aircraft types, determining a performance envelope that defines the performance of a weapon when it is implemented on each of the different aircraft types. In one example, the performance envelope is the smallest-sized envelope that defines the performance of a weapon when it is implemented on each of the different aircraft types.

[0079] In some embodiments, the database is generated by defining the range of conditions under which a weapon may be required to be fired, the range of aircraft conditions under which an aircraft can fire a weapon, and the range of weapon conditions under which a weapon can fire; generating data showing weapon performance for each weapon firing possibility from within the defined ranges; and creating a database that defines the overall performance envelope of the weapon. Coefficients can then be determined from this database and a general polynomial. In this way, the database can be generated on a ground-based system, and therefore the aircraft system only needs capacity to store a general polynomial and process the coefficients using aircraft and target conditions in order to generate the feasibility representation. As a result, the amount of data storage / processing capacity required on the aircraft tends to be reduced.

[0080] The coefficients can be implemented as loadable data to enable accurate and precise weapon behavior within the weapon system. Furthermore, using only one or a few common algorithms would allow different weapon systems to be authorized or certified / approved for use with aircraft with reduced effort and more quickly than the extensive testing required by conventional approaches.

[0081] The step of uploading the generated coefficients to the aircraft can be performed when the weapon is loaded as an aircraft stowage. When loading a new weapon stowage, in order to integrate the weapon with the aircraft aiming system, the coefficients associated with the weapon can be uploaded to the aircraft simultaneously with the weapon. Preferably, the coefficients are stored on a hardware device together with the weapon, and the device is connected to the aircraft to upload the coefficient data when the weapon is loaded.

[0082] In one example, the types of candidate polynomials in the set of candidate polynomials include univariate polynomials, multivariate polynomials, and modified forms thereof. Other polynomial types are known.

[0083] In one example, the degree of the candidate polynomials in the set of candidate polynomials is in the range of 1 to 100, preferably in the range of 2 to 25, more preferably in the range of 3 to 10, and most preferably in the range of 5 to 9, such as 5, 6, 7, 8, 9.

[0084] In one example, a general polynomial is in the following form:

[0085]

Number

[0086] Here: α mn represents the m coefficients required to calculate the output n, {x1…x Ni} represents the normalized input, {y1…y Ni} represents the output, p 1mn represents the power (exponent) of the x1 variable of the m-th term of the n-th polynomial.

[0087] In one example, the best candidate polynomial is in the following form:

[0088]

Number

[0089] Here:

[0090]

number

[0091]

number

[0092]

number

[0093] In one example, the order of a general polynomial is 3 or greater. In another example, the order of a general polynomial is in the range of 10 to 25, for example, 20. Surprisingly, the inventors have found that using a general algorithm with an order of about 20 accurately and sufficiently describes most air-to-air engagements at a suitable runtime for an on-air implementation. Nevertheless, general algorithms can have an order greater than 2.

[0094] In one example, step b) calculates the coefficients for each candidate polynomial that best fits that candidate polynomial to the characteristics of the weapon's performance envelope, using the least-squares error criterion, and comprises: 1) generating an initial population of candidate polynomials; 2) for each candidate polynomial, calculating a set of coefficients that fits that polynomial to the performance envelope according to one or more criteria; 3) for each set of candidate polynomials and coefficients, calculating a score function indicating the quality of the fit of that set of candidate polynomials and coefficients to the performance envelope; and 4) recursively applying a genetic algorithm to the set of candidate polynomials until one or more criteria are met, including retaining at least the best-scoring polynomials and discarding the other polynomials. In one example, the output of the retained polynomials is a layer of a self-organizing polynomial neural network, which is used to provide input for creating higher-order candidate polynomials. In one example, these steps are iterated over the higher-order candidate polynomials. In one example, the final result is obtained from the path that ends with the best candidate score.

[0095] In one example, the target is equipped with and / or is an aircraft. In another example, the feasibility indication shows the successful launch zone for the aircraft and / or the target.

[0096] In one example, the target includes and / or is a ground-based target. In another example, the feasibility indication shows the aircraft's launch area and / or the target's missile engagement zone.

[0097] For example, step b) calculates the coefficients for each candidate polynomial that best fits the characteristics of the weapon's performance envelope, using the least-squares error criterion. The process involves generating an initial population of candidate polynomials, For each candidate polynomial, calculate a set of coefficients that fit the polynomial to the performance envelope according to one or more criteria (e.g., the least squares criterion), For each set of candidate polynomials and coefficients, calculate a score function that indicates the quality of the fit of that set of candidate polynomials and coefficients to the performance envelope. The genetic algorithm is applied recursively to the set of candidate polynomials until one or more criteria are met, including retaining at least the best-scoring polynomial and discarding other polynomials. It is equipped with.

[0098] In one example, the output of the retained polynomial(s) is a layer of a self-organizing polynomial neural network, used to provide input for constructing higher-order candidate polynomials. In one example, these are iterated over until the final result with the best candidate score is obtained.

[0099] In one example, the performance envelope of a weapon is the minimum envelope that defines the performance of the weapon when it is mounted on an aircraft, e.g., the minimum envelope that defines the performance of the weapon. In another example, the performance envelope of a weapon is the performance of the weapon when it is mounted on an aircraft of a different aircraft type. In another example, the method comprises obtaining the respective performance envelopes for one or more different aircraft types, e.g., multiple different aircraft types.

[0100] In one example, the method comprises determining a performance envelope using multiple aircraft performance envelopes, including determining a performance envelope that defines all the performance of different aircraft types (i.e., a "maximum aircraft performance envelope"), and using the performance envelopes and weapon performance envelopes representing all the performance of different aircraft types, determining a performance envelope that defines the performance of a weapon when it is implemented on each of the different aircraft types. In one example, the performance envelope is the smallest-sized envelope that defines the performance of a weapon when it is implemented on each of the different aircraft types.

[0101] In some embodiments, the database is generated by defining the range of conditions under which a weapon may be required to be fired, the range of aircraft conditions under which an aircraft can fire a weapon, and the range of weapon conditions under which a weapon can fire; generating data showing weapon performance for each weapon firing possibility from within the defined ranges; and creating a database that defines the overall performance envelope of the weapon. Coefficients can then be determined from this database and a general polynomial. In this way, the database can be generated on a ground-based system, and therefore the aircraft system only needs capacity to store a general polynomial and process the coefficients using aircraft and target conditions in order to generate the feasibility representation. As a result, the amount of data storage / processing capacity required on the aircraft tends to be reduced.

[0102] The coefficients can be implemented as loadable data to enable accurate and precise weapon behavior within the weapon system. Furthermore, using only one or a few common algorithms would allow different weapon systems to be authorized or certified / approved for use with aircraft with reduced effort and more quickly than the extensive testing required by conventional approaches.

[0103] The step of uploading the generated coefficients to the aircraft may be performed when the weapon is loaded as an aircraft containment. When a new weapon containment is loaded, the coefficients associated with that weapon may be uploaded to the aircraft at the same time as the weapon in order to integrate the weapon with the aircraft targeting system. Preferably, the coefficients are stored on a hardware device along with the weapon, and the device is connected to the aircraft to upload the coefficient data when the weapon is loaded.

[0104] According to a second aspect of the present invention, a weapon carried on an aircraft targets a target Successfully engage in combat The feasibility and / or the weapons that can be delivered on the target are with aircraft and Successfully engage in combatA system is provided for generating a feasibility statement in an aircraft in flight, the system comprising a first computer having memory and a processor and located away from the aircraft, and a second computer having memory and a processor and located on board the aircraft. The first computer was, To provide a database that describes the performance envelope of weapons, The process involves creating the coefficient characteristics of its performance envelope using a general algorithm, which includes a step of identifying the best candidate polynomial from several candidate polynomials, wherein the general algorithm has the form of a polynomial, and the variables of the polynomial are some or all of a group of firing condition parameters for a weapon or aircraft. Upload the coefficients of the best identified candidate polynomial to the second computer. The second computer is configured to do the following: A reconfigurator containing the same general algorithm selects coefficients for the general algorithm according to the aircraft and target conditions, Using the selected coefficients, the reconstructor generates a feasibility representation. In a system configured to perform the above, the second computer is configured to use a reconfigurator containing the same general algorithm to select coefficients for the general algorithm according to the conditions of the aircraft and target, when the aircraft and target are within the performance envelope of the weapon.

[0105] In one example, the system includes a display for showing feasibility.

[0106] According to a third aspect of the concept of the present invention, an aircraft is provided which is equipped with a second computer according to the second aspect.

[0107] A fourth aspect of the present invention provides a computer comprising a processor and memory, configured to carry out the method according to the first aspect.

[0108] According to a fifth aspect of the present invention, a computer program is provided which, when executed by a computer having a processor and memory, provides instructions for causing the computer to perform the method according to the first aspect.

[0109] According to a sixth aspect of the present invention, a non-transient computer-readable storage medium is provided which, when executed by a computer having a processor and memory, provides instructions for causing the computer to perform the method according to the first aspect.

[0110] Herein, embodiments of the present invention will be described only by reference to the drawings. [Brief explanation of the drawing]

[0111] [Figure 1A] A schematic diagram of the firing range (LAR) for air-to-ground weapons is shown. [Figure 1B] A schematic diagram of the firing range (LAR) for air-to-ground weapons is shown. [Figure 2] A schematic diagram of the successful launch zone (LSZ) for air-to-air weapons is shown. [Figure 3] A schematic diagram of a system according to an illustrative embodiment is provided. [Figure 4] The system shown in Figure 3 is illustrated in more detail. [Figure 5] Figure 3 provides a more detailed schematic diagram of the system, illustrating the configuration of SOPNN. [Figure 6] Figure 3 provides a more detailed schematic illustration of the system, showing the application of the GA-SOPNN model for classifying weapon capabilities and estimating LSZ / LAR parameters in the domain of engagement. [Figure 7] Figure 3 provides a more detailed schematic diagram of the system, illustrating a polynomial neural network-based capability filter. [Modes for carrying out the invention]

[0112] Figure 1A schematically illustrates the LAR in the flight plane of a launch aircraft 1 flying along flight path 3 toward a target 5 for an air-to-ground weapon (not shown) mounted on the aircraft. The LAR is calculated to provide a cockpit display in the launch aircraft 1 regarding the feasibility of the situation and the opportunity to fire. Figure 1B schematically illustrates the display generated for the LAR in Figure 1A, which is in the form of a downrange and crossrange display (shaded area), where the weapon flight path 7 coincides with the aircraft flight path 3 and is toward target 5 as shown in the display. Successfully engage in combat For this to happen, the target must enter the shaded LAR. When aircraft 1 moves in the downrange direction, the displayed LAR is the minimum range R. min and maximum range R max It is bounded by.

[0113] In addition to the LAR for launch aircraft 1, the Missile Engagement Zone (MEZ) for target 5 is determined and can be displayed to the pilot of aircraft 1. This MEZ is the area where surface-to-air weapons (e.g., missiles) carried by target 5 can be launched from aircraft 1. They successfully intercepted it. This may indicate a region where the probability exceeds a threshold.

[0114] The LSZ shown in Figure 2 is the region where the probability of an air-to-air weapon hitting an aerial target T exceeds a threshold level. Calculating the LSZ is more complex than for the LAR because it involves a greater number of factors, including the relative speed and direction of the launching aircraft and the target, as well as the relative speed and direction of the weapon relative to the target. Furthermore, the shape of the LSZ is more complex than that of the LAR, and similar to the LAR, the probability of target T hitting within it is higher. Successfully engage in combat The maximum range R that can be reached min and minimum range R max However, considering the speed and direction of the launching aircraft and target T, it is outside the range of the launching aircraft's ability to operate the weapon and hit the target when it is very close to the target, therefore target T Successfully engage in combat It is not possible, R min There is a zone bounded by R. In this example, LSZ is the so-called "inescapable zone" R NE R NEand R min The zone demarcated by is where target T is weaponized. Successfully avoided This is the zone where the probability falls below a threshold probability. This range can be determined using the performance parameters of the weapon, the launch aircraft 1, and the target T. As is known in the art, there are two LSZs, one for the launch aircraft to engage target T and the other for the target to engage the launch aircraft.

[0115] In many cases, it is necessary to calculate the LAR or LSZ for an engagement in order to display information to the launch aircraft crew regarding the feasibility or likelihood of success of the engagement, and to assist in firing control and guidance decisions. The conventional approach is to create a simple abstract model of the weapon with parameters defined by the launch conditions, which is then mounted on the launch aircraft and used to generate the LAR, LSZ, or MEZ and appropriate representations.

[0116] Traditional weapon targeting processes typically involve custom design, implementation, and certification for every weapon / platform combination. This custom approach is extremely costly and time-intensive. Even slight variations in weapon performance trigger complex loops around the custom process. Such processes have also tended to have limited computing capacity. However, today's platforms exceed critical capacity limits. To improve weapon integration time and affordability, the inventors have developed a Data Driven Weapon Integration (DDWI) approach that uses a general algorithm that can be “customized” for a specific weapon by using a unique set of data coefficients. In doing so, DDWI breaks the reliance on custom models for weapon targeting. It reduces the lifecycle cost of avionics equipment, improves the timescale, resulting in significant savings, and provides flexibility in services for “tuning” and / or “sanitizing” weapon targeting performance. Mission data coefficients uploaded to the platform are derived from advanced multidimensional weapon models. By performing parallel computing on multi-core computers, GPUs, and computer clusters, the inventors solve such computational and data-intensive problems, freeing up more performance and processing time reductions.

[0117] Figure 3 schematically illustrates the system according to an illustrative embodiment. The DDWI has three elements: • Adaptation toolset, office-based design tools for mission systems engineers to translate weapon truth data into a set of weapon coefficients. • Data load coefficient, a common format which may be software build / theater / mission data loaded in different platforms for the same weapon, and A common onboard algorithm that can represent any weapon on any platform and is weaponized only when a coefficient is loaded.

[0118] Figure 4 provides a more detailed schematic diagram of the system shown in Figure 3, divided into process 11, which is carried out on the ground, and process 13, which is carried out on the launch aircraft 1. The system involves a weapon delivered on an aircraft to a target. Successfully engage in combat The feasibility and / or the weapons that can be delivered on the target are with aircraft and Successfully engage in combat The system is for generating a feasibility statement indicating feasibility while in flight, and comprises a first computer having memory and a processor, located away from the aircraft, and a second computer having memory and a processor, mounted on aircraft 1.

[0119] The first computer 11 is configured to provide a database describing the performance envelope of a weapon, and to create the coefficient characteristics of that performance envelope using a general algorithm, by steps including identifying the best candidate polynomial from a plurality of candidate polynomials, wherein the general algorithm has the form of a polynomial, and the variables of the polynomial are some or all of a group of firing condition parameters of a weapon or aircraft, and to upload the coefficients of the identified best candidate polynomial to the second computer.

[0120] The second computer 13 is configured to use a reconfigurator containing the same general algorithm to select coefficients for the general algorithm according to the conditions of the aircraft and target, and to use the selected coefficients to generate a feasibility representation. The second computer 13 is configured to use a reconfigurator containing the same general algorithm to select coefficients for the general algorithm according to the conditions of the aircraft and target, when the aircraft and target are within the performance envelope of the weapon.

[0121] In this way, a capability filter is provided to determine the limits of the possible envelopes of weapon systems within an arbitrarily selected region of the engagement envelope and to examine the feasibility of weapon engagement under current firing conditions, i.e., to determine whether the current firing is inside or outside the "zone of impact" for the weapon. This offers the prospect of classifying large, high-dimensional spaces using a relatively simple model, thereby saving both processing and memory for the host system.

[0122] Figure 6 illustrates how the capability filter fits into the assessment of the entire envelope. This figure represents the process performed for a candidate engagement. First, the capability filter assesses whether the weapon is capable or not. Second, if the weapon is capable, the relevant LSZ / LAR parameters are estimated.

[0123] Figure 7 outlines the three main steps involved in CF implementation. First, all capability ranges in the training data are converted to 1, and no capability cases are kept at 0. Next, the binary output is numerically learned using the fitting and estimation methods described above. Finally, a threshold is applied to the predicted value (somewhere between 1 and 0) to determine the true binary output.

[0124] More specifically, the core of DDWI is the offline coefficient generator 21. The coefficient generator 21 identifies coefficients for a general algorithm to "fit" the general algorithm to the performance envelope linearity. The form of the general algorithm is usually predetermined, for example, an arbitrary polynomial of degree (i.e., order) up to n. The coefficient generator 21 receives the true performance envelope and calculates coefficients for the general algorithm. The coefficients "fit" the general algorithm to the performance envelope linearity.

[0125] The estimation and fitting process uses a genetic algorithm for the self-organizing neural network approach. It computes a set of coefficients that will allow the geometric shape of the LAR / LSZ region to be modeled (and subsequently reconstructed) by a standard polynomial "algorithm" (see Figure 5). It employs an evolutionary technique called the genetic algorithm as a central mechanism for the self-organizing polynomial neural network (GA-SOPNN), automating the derivation of coefficients for several polynomial models within each layer. This process involves the following steps: 1. Create an initial population of candidate polynomials of different orders, each having an input that contains some or all of the firing parameters, where each polynomial function is a unique solution to the problem. 2. Using the least squares error criterion, calculate coefficients for fitting these candidates to the weapon performance envelope for selected LAR / LSZ characteristics. 3. Calculate the score function for each candidate, 4. Recursively improve this population using a genetic algorithm, a) Keep the best score candidates, b) Reject the worst candidate, c) "Grow" a new population with randomly selected combinations of features extracted from the best group. 5. Repeat until improvement stops or the accuracy criteria are met. 6. The result is the first layer of a self-organizing polynomial neural network (SOPNN), 7. Each subsequent layer of SOPNN takes the best output of the previous layer as its input, and then proceeds as described above. 8. The effect is to create higher-order candidate polynomials for consideration, 9. Optimization within the new layer uses the same genetic algorithm as before. 10. Layers will be added until improvement stops or the maximum layer set by the user is reached. 11. Only the single best polynomial and coefficient set of interest, i.e., the output of the final layer with the best score. 12. All other outputs of the final layer are rejected. 13. All nodes in the lower layers that do not contribute to the best polynomial are rejected.

[0126] To apply a genetic algorithm to a specific purpose, an internal representation of the space to be explored is selected, and an external function is defined that assigns goodness-of-fit values ​​to candidate solutions.

[0127] In this example, the method comprises using a trained machine learning model, such as a trained neural network, to infer whether the aircraft and target are within the performance envelope of the weapon, according to the conditions of the aircraft and target.

[0128] In this example, the method involves training a machine learning model using training data of the performance envelopes of each weapon, according to the conditions of each aircraft and each target.

[0129] In this example, the method comprises labeling training data based on whether each aircraft and each target are within the performance envelope of their respective weapons, according to the conditions of each aircraft and each target.

[0130] In this example, the method comprises creating the coefficient characteristics of each performance envelope using a general algorithm, by stepping in a step that includes identifying the best candidate polynomial from several candidate polynomials, where the variables of the polynomials are some or all of the group of firing condition parameters for each weapon or aircraft.

[0131] In this example, inferring whether an aircraft and a target are within the performance envelope of a weapon using a trained machine learning model, according to the conditions of the aircraft and the target, involves thresholding the results of the inference.

[0132] In this example, the coefficient characteristics of its performance envelope are created using a general algorithm, which involves a step of identifying the best candidate polynomial from several candidate polynomials, wherein the general algorithm has the form of a polynomial, and the variables of the polynomial are some or all of a group of firing condition parameters for a weapon or aircraft. a) Generate candidate polynomials, where the variables of the polynomials are some or all of the group of firing condition parameters for weapons or aircraft. b) For each candidate polynomial, use the least-squares error criterion to calculate the coefficients for the candidate polynomial that best fits the characteristics of the weapon's performance envelope, c) For each candidate polynomial, a candidate score is generated according to the quality of the fit of that candidate polynomial to the characteristics of the weapon's performance envelope, d) Apply a genetic algorithm to the candidate polynomials and scores, including selecting the best-scoring polynomial(s) and discarding the other polynomial(s), thereby identifying the best candidate polynomial and its coefficients. e) Repeat the identification process until all required properties of the performance envelope have corresponding polynomial models. The process involves a step that includes creating the coefficient characteristics of its performance envelope using a general algorithm, which has the form of a polynomial.

[0133] More specifically, processes 11 and 13 begin with the generation of a data space which is the range of conditions under which the weapon performance envelope should be defined, which is done by the data space generator 15 and depends on the range of conditions under which it is necessary to fire the weapon (defined by the weapon user / operator), is feasible to fire according to the launch aircraft capability, and is feasible to fire according to the weapon capability / performance.

[0134] In this example, the data space generator 15 contains data describing performance parameters for each of several different aircraft types. Different types of aircraft may have different capabilities from one another; for example, aircraft with the same or similar capabilities may be considered the same “aircraft type”. Different types of aircraft may be different models or variants and / or may have different manufacturers. Different types of aircraft may have different operating parameters (maximum speed, maximum altitude, g limit, etc.). Different types of aircraft may be configured for different purposes or functions (e.g., bombers, fighters, refueling aircraft, etc.). These aircraft performance envelopes may be supplied by the aircraft manufacturer or through testing. Several different aircraft types include the type of launch aircraft 1 and preferably the type of target aircraft T. Performance parameters for each aircraft type may include, but are not limited to, the maximum achievable altitude, the maximum achievable g force, and the maximum achievable climb angle. The values ​​of the performance parameters for different types of aircraft may differ from one another. For example, a first type of aircraft may have a maximum altitude of 45,000 feet, while a second type of aircraft may have a maximum altitude of 55,000 feet.

[0135] In this example, the data space generator 15 further includes data describing performance parameters for each of several different weapon types, for example, different weapons that may be loaded onto a launch aircraft or that are expected to be carried by an adversarial target. These weapon performance envelopes may be supplied by the weapon manufacturer or through testing. The several different weapon types include the type of weapon carried by the launch aircraft 1, and preferably the target. The performance parameters for each weapon type may include, but are not limited to, the maximum altitude from which the weapon can be launched, the maximum g-force from which the weapon can be launched, and the weapon's release mechanism. The values ​​of the performance parameters for different types of weapons may differ from one another. For example, a first type of weapon may be capable of being launched to an altitude of 35,000 feet, while a second type of weapon may be capable of being launched to an altitude of 45,000 feet, and so on.

[0136] The data space generator 15 can define emission, weather, and commanded impact conditions for training and validation sets performed by the truth data generator 17.

[0137] The truth data generator 17 determines the weapon performance for each firing case in the data space, which depends on the weapon performance model typically provided by the weapon manufacturer.

[0138] The output of the truth data generator 17 is the truth database 19, which is a set of data specifying further weapon performance envelopes for each of several exemplary weapon firings for each weapon type. The truth data generator 17 can produce training and validation sets used by the coefficient generator 21.

[0139] Traditionally, the Truth Database has been used as a model that can be mounted on a launching aircraft to generate the feasibility of engagement indications (LAR or LSZ, as appropriate).

[0140] In this example, the coefficient generator 21 receives additional weapon performance envelopes stored in the truth database 19 and calculates coefficients for each weapon type and each exemplary weapon firing according to a general LAR / LSZ algorithm 23 that "fits" a general algorithm to the additional weapon performance envelope linearity.

[0141] This describes a method for determining coefficient values ​​that fit a general algorithm to the performance envelope of a specific weapon type and a specific example of weapon firing. In practice, it will be recognized that a set of coefficients is determined for each example of weapon firing and for each weapon type.

[0142] In this method, the coefficient generator 21 begins by creating an initial set of candidate polynomials whose variables are some or all of the weapon or aircraft firing condition parameters. Each candidate polynomial is a unique solution to the fitting problem. Some or all of the candidate polynomials may have a different order or dimension than some or all of the other candidate polynomials. For each candidate polynomial, the set of coefficients that best "fit" that candidate polynomial to the weapon performance envelope is then calculated. This can be done using the least-squares error criterion or any other fitting method. For each candidate polynomial, a score indicating the quality of this fit is then calculated.

[0143] The number of inputs (27) and the format of each polynomial descriptor, as well as the PD layer nodes, are determined by an optimization method known as the genetic algorithm. The genetic algorithm is applied to candidate polynomials and scores. In this example, the polynomial with the best score is retained, and other polynomials (i.e., those with the worst score) are rejected. New candidate polynomials with similar characteristics to the retained candidate polynomials are then created and replaced (for example, by "multiplying" and "mutating" the retained candidate polynomials). For the generation of these new candidates, a set of coefficients and score values ​​is then calculated, and so on.

[0144] The genetic algorithm is repeated until the best candidate's score is improved or some other criterion is met. The result is the first layer of the self-organizing polynomial neural network (SOPNN), i.e., layer 1.

[0145] The entire process is then repeated, with the output of the first layer providing the input to create the second layer of the SOPNN, i.e., layer 2. The new layer has the effect of generating higher-order candidate polynomials and coefficients for consideration. The selection of polynomials in the new layer is again governed and optimized by a genetic algorithm.

[0146] Layers are added to the SOPNN in this way until the score of the best candidate is improved or some other criterion is met. The completed network with two layers is shown in Figure 5. The final network is obtained recursively from paths ending with the output node that has the best score in the final generation of candidates ("optimal solution"). Any node that does not have a connection to this path is discarded as shown in Figure 5, where nodes that contribute to the optimal solution are lightly shaded and discarded nodes are black.

[0147] As described above, the inventors have adapted the genetic algorithm to be parallelized across the many polynomial orders and input parameters that the genetic algorithm must perform.

[0148] The best single candidate polynomial and coefficient set is identified and stored. This process is repeated until all required properties of the LAR / LSZ have corresponding polynomial models. In other words, this process is repeated for each firing condition and each weapon type until a polynomial model is generated that fits further weapon performance envelopes for that weapon type and firing condition.

[0149] The general LAR / LSZ algorithm is predetermined, and in this example, the general polynomial is of the following form, as previously described:

[0150]

number

[0151] In this example, the best candidate polynomial is of the following form, as previously explained:

[0152]

number

[0153] In this example, the order of a general polynomial is in the range of 10 to 25, for example, 20.

[0154] Referring again to Figure 4, the output of the coefficient generator 21 is a set of coefficients that are loaded into the launching aircraft by the data uploader. Following this step, the onboard process 13 includes a reconfigurator 25, which combines the uploaded coefficients with a general LAR / LSZ algorithm 23 (held in the aircraft system) to reconfigure the LAR, LSZ, or MEZ for a particular engagement by selecting appropriate algorithms and coefficients for the current firing conditions (i.e., weapon or aircraft firing conditions).

[0155] When a LAR, LSZ, or MEZ is reconfigured for a specific engagement by a system mounted on an aircraft, the LAR, LSZ, or MEZ is displayed by conventional means mounted on the aircraft. In this example, when launch aircraft 1 engages an adversarial target aircraft T during operation, the reconfigurator 25 mounted on launch aircraft 1 may select from uploaded coefficients that correspond to the weapon carried by launch aircraft 1 and to the relevant firing conditions (experienced altitude, angle of attack, environmental conditions, g-force, etc.). The selected coefficients may then be used to reconfigure the LSZ of launch aircraft 1 and display it to the pilot of launch aircraft 1. The reconfigured LSZ of launch aircraft 1 may also be used by other systems mounted on launch aircraft 1 to recommend action to the pilot of launch aircraft 1 (e.g., a recommendation to fire the weapon, etc.).

[0156] When launch aircraft 1 engages an adversarial target aircraft T, the aircraft type of the adversarial target T may be determined by the pilot of launch aircraft 1 (or by other means) and entered into the reconfigurator 25. The reconfigurator 25 mounted on launch aircraft 1 may then select from the uploaded coefficients the coefficients corresponding to the weapon most likely to be carried by the adversarial target T and the associated firing conditions. The selected coefficients may then be used to reconfigure the LSZ of the adversarial target T and display it to the pilot of launch aircraft 1. The reconfigured LSZ of the adversarial target T may also be used by other systems mounted on launch aircraft 1 to recommend actions to the pilot of launch aircraft 1 (e.g., a recommendation that a particular evasive maneuver be performed).

[0157] In this example, when the launch aircraft 1 engages a hostile ground target 5 during operation, the reconfigurator 25 mounted on the launch aircraft 1 may select from uploaded coefficients that correspond to the weapon carried by the launch aircraft 1 and to the relevant firing conditions (experienced altitude, angle of attack, environmental conditions, g-force, etc.). The selected coefficients may then be used to reconfigure the LAR of the launch aircraft 1 and display it to the pilot of the launch aircraft 1. The reconfigured LAR of the launch aircraft 1 may also be used by other systems mounted on the launch aircraft 1 to recommend actions to the pilot of the launch aircraft 1 (e.g., a recommendation to fire the weapon, etc.).

[0158] When launch aircraft 1 engages a hostile ground target 5, the type of ground target 5 may be determined by the pilot of launch aircraft 1 (or by other means) and entered into the reconfigurator 25. The reconfigurator 25 mounted on launch aircraft 1 may then select from the uploaded coefficients the coefficient corresponding to the weapon most likely to be carried by ground target 5 and the associated firing conditions. The selected coefficient may then be used to reconfigure the MEZ of ground target 5 and display it to the pilot of launch aircraft 1. The reconfigured MEZ of ground target 5 may also be used by other systems mounted on launch aircraft 1 to recommend actions to the pilot of launch aircraft 1 (e.g., a recommendation that a particular evasive maneuver be performed).

[0159] An apparatus for implementing the configurations described above, including any of the processors mentioned above, may be provided by configuring or adapting any suitable apparatus, for example, one or more computers or other processing units or processors, and / or by providing additional modules. The apparatus may comprise a computer, a network of computers, or one or more processors for implementing instructions and using data, including instructions and data in the form of computer programs or multiple computer programs stored in or on machine-readable storage media such as computer memory, computer disks, ROM, PROM, etc., or any combination thereof or other storage media. The following is a direct reproduction of the claims as originally filed. [1] A computer implementation method for generating a feasibility representation indicating the feasibility of a weapon carried on an aircraft in flight successfully engaging a target and / or the feasibility of a weapon carried on a target successfully engaging the aircraft, the method comprising providing a database describing the performance envelope of the weapon, The process involves creating the coefficient characteristics of its performance envelope using a general algorithm, which includes a step of identifying the best candidate polynomial from several candidate polynomials, wherein the general algorithm has the form of a polynomial, and the variables of the polynomial are some or all of a group of firing condition parameters for a weapon or aircraft. Uploading the coefficients of the identified best candidate polynomial to the aircraft; selecting the coefficients for the general algorithm according to the conditions of the aircraft and the target by a reconstructor on the aircraft that includes the same general algorithm; Using the selected coefficients, the reconstructor generates the feasibility representation. A method comprising the following, wherein the reconfigurator on the aircraft, which includes the same general algorithm, selects coefficients for the general algorithm according to the conditions of the aircraft and the target, and the reconfigurator on the aircraft, which includes the same general algorithm, selects coefficients for the general algorithm according to the conditions of the aircraft and the target when the aircraft and the target are within the performance envelope of the weapon. [2] The method of [1], comprising using a trained machine learning model to infer whether the aircraft and the target are within the performance envelope of the weapon, according to the conditions of the aircraft and the target. [3] The method according to [2], comprising training a machine learning model using training data of the performance envelope of each weapon in accordance with the conditions of each aircraft and each target. [4] The method according to [3], further comprising labeling the training data based on whether each of the aircraft and each of the targets lies within the performance envelope of the respective weapon, according to the conditions of each of the aircraft and each of the targets. [5] The method according to [3] or [4], further comprising creating the coefficient characteristics of each of the performance envelopes using the general algorithm, by a step of identifying the best candidate polynomial from a plurality of candidate polynomials, wherein the variables of the polynomials are some or all of a group of firing condition parameters for each weapon or aircraft. [6] The method of any one of [2] to [5], wherein the method comprises using the machine learning model trained in accordance with the conditions of the aircraft and the target to infer whether the aircraft and the target are within the performance envelope of the weapon, and thresholding the result of the inference. [7] The method according to any one of [1] to [6], wherein the reconfigurator on the aircraft, which includes the same general algorithm, selects coefficients for the general algorithm according to the conditions of the aircraft and the target when the aircraft and the target are within the performance envelope of the weapon, and the reconfigurator on the aircraft, which includes the same general algorithm, selects coefficients for the general algorithm when the aircraft and the target are currently within the performance envelope of the weapon. [8] The method according to any one of [1] to [7], wherein the reconfigurator on the aircraft, which includes the same general algorithm, selects coefficients for the general algorithm according to the conditions of the aircraft and the target when the aircraft and the target are within the performance envelope of the weapon, and the reconfigurator on the aircraft, which includes the same general algorithm, selects coefficients for the general algorithm only when the aircraft and the target are within the performance envelope of the weapon. [9] The method according to any one of [1] to [8], wherein the reconfigurator on the aircraft, which includes the same general algorithm, selects coefficients for the general algorithm according to the conditions of the aircraft and the target when the aircraft and the target are within the performance envelope of the weapon, and the reconfigurator on the aircraft, which includes the same general algorithm, selects coefficients for the general algorithm while the aircraft and the target are within the performance envelope of the weapon.

[10] The method according to any one of [1] to [9], wherein the reconfigurator on the aircraft, which includes the same general algorithm, selects coefficients for the general algorithm according to the conditions of the aircraft and the target when the aircraft and the target are within the performance envelope of the weapon, and further comprises the reconfigurator on the aircraft, which includes the same general algorithm, repeatedly selecting coefficients for the general algorithm when the aircraft and the target are within the performance envelope of the weapon.

[11] The method according to any one of [1] to

[10] , wherein the reconfigurator on the aircraft, which includes the same general algorithm, selects coefficients for the general algorithm according to the conditions of the aircraft and the target when the aircraft and the target are within the performance envelope of the weapon, and further comprises the reconfigurator on the aircraft, which includes the same general algorithm, deselecting coefficients for the general algorithm when the aircraft and the target are no longer within the performance envelope of the weapon.

[12] A system for generating feasibility indicators in flight of an aircraft indicating the feasibility of a weapon carried on an aircraft successfully engaging a target and / or the feasibility of a weapon carried on the target successfully engaging the aircraft, the system comprising a first computer having a computer, i.e., memory and a processor, and being located away from the aircraft, and a second computer having a computer having a computer, i.e., memory and a processor, and being mounted on the aircraft, The aforementioned first computer, To provide a database describing the performance envelope of the aforementioned weapon, The process involves creating the coefficient characteristics of its performance envelope using a general algorithm, which includes a step of identifying the best candidate polynomial from several candidate polynomials, wherein the general algorithm has the form of a polynomial, and the variables of the polynomial are some or all of a group of firing condition parameters for a weapon or aircraft. Uploading the coefficients of the identified best candidate polynomial to the second computer. The second computer is configured to perform the following: A reconfigurator containing the same general algorithm selects coefficients for the general algorithm according to the conditions of the aircraft and the target, Using the selected coefficients, the reconstructor generates the feasibility representation. A system configured to perform the following, wherein the second computer is configured to use the reconfigurator, which includes the same general algorithm, to select coefficients for the general algorithm according to the conditions of the aircraft and the target, when the aircraft and the target are within the performance envelope of the weapon.

[13] The system according to

[12] , further comprising a display for displaying the feasibility indication.

[14] An aircraft equipped with a second computer as described in

[12] or

[13] .

[15] A computer having a processor and memory configured to perform the method described in any one of [1] to

[11] , a computer program having instructions that, when executed by the computer having a processor and memory, cause the computer to perform the method described in any one of [1] to

[11] , or a non-transient computer-readable storage medium having instructions that, when executed by the computer having a processor and memory, cause the computer to perform the method described in any one of [1] to

[11] .

Claims

1. A computer implementation method for generating a feasibility display indicating the feasibility of a weapon carried on an aircraft successfully engaging a target and / or the feasibility of a weapon carried on a target successfully engaging the aircraft, wherein the method is: To provide a database describing the performance envelope of the aforementioned weapon, The process involves creating the coefficient characteristics of its performance envelope using a general algorithm, which includes a step of identifying the best candidate polynomial from several candidate polynomials, wherein the general algorithm has the form of a polynomial, and the variables of the polynomial are some or all of a group of firing condition parameters for a weapon or aircraft. Uploading the coefficients of the identified best candidate polynomial to the aircraft, A reconfigurator on the aircraft, which includes the same general algorithm, selects coefficients for the general algorithm according to the conditions of the aircraft and the target, Using the selected coefficients, the reconstructor generates the feasibility representation. A method comprising, wherein the reconfigurator on the aircraft, which includes the same general algorithm, selects coefficients for the general algorithm according to the conditions of the aircraft and the target, when the aircraft and the target are within the performance envelope of the weapon.

2. The method according to claim 1, comprising using a trained machine learning model to infer whether the aircraft and the target are within the performance envelope of the weapon, in accordance with the conditions of the aircraft and the target.

3. The method according to claim 2, comprising training the machine learning model using training data of the performance envelope of each weapon according to the conditions of each aircraft and each target.

4. The method according to claim 3, comprising labeling the training data based on whether each of the aircraft and each of the targets is within the performance envelope of the respective weapon, according to the conditions of each of the aircraft and each of the targets.

5. The method according to claim 3, comprising the step of creating the coefficient characteristics of each of the performance envelopes using the general algorithm, wherein the variables of the polynomials are some or all of a group of firing condition parameters for each weapon or aircraft.

6. The method of claim 2, wherein, according to the conditions of the aircraft and the target, using the trained machine learning model, inferring whether the aircraft and the target are within the performance envelope of the weapon, the result of which a binary output is determined.

7. The method according to claim 1, wherein the reconfigurator on the aircraft, which includes the same general algorithm, selects coefficients for the general algorithm according to the conditions of the aircraft and the target when the aircraft and the target are within the performance envelope of the weapon, and the reconfigurator on the aircraft, which includes the same general algorithm, selects coefficients for the general algorithm when the aircraft and the target are currently within the performance envelope of the weapon.

8. The method according to claim 1, wherein the reconfigurator on the aircraft, which includes the same general algorithm, selects coefficients for the general algorithm according to the conditions of the aircraft and the target when the aircraft and the target are within the performance envelope of the weapon, and the reconfigurator on the aircraft, which includes the same general algorithm, selects coefficients for the general algorithm only when the aircraft and the target are within the performance envelope of the weapon.

9. The method according to claim 1, wherein the reconfigurator on the aircraft, which includes the same general algorithm, selects coefficients for the general algorithm according to the conditions of the aircraft and the target when the aircraft and the target are within the performance envelope of the weapon, and the reconfigurator on the aircraft, which includes the same general algorithm, selects coefficients for the general algorithm while the aircraft and the target are within the performance envelope of the weapon.

10. The method according to claim 1, wherein the reconfigurator on the aircraft, which includes the same general algorithm, selects coefficients for the general algorithm according to the conditions of the aircraft and the target when the aircraft and the target are within the performance envelope of the weapon, and further comprises repeatedly selecting coefficients for the general algorithm when the aircraft and the target are within the performance envelope of the weapon, by the reconfigurator on the aircraft, which includes the same general algorithm.

11. The method according to claim 1, wherein the reconfigurator on the aircraft, which includes the same general algorithm, selects coefficients for the general algorithm according to the conditions of the aircraft and the target when the aircraft and the target are within the performance envelope of the weapon, and the reconfigurator on the aircraft, which includes the same general algorithm, deselects coefficients for the general algorithm when the aircraft and the target are no longer within the performance envelope of the weapon.

12. A system for generating a feasibility indicator in a flying aircraft showing the feasibility of a weapon carried on the aircraft successfully engaging a target and / or the feasibility of a weapon carried on the target successfully engaging the aircraft, the system comprising a first computer having a computer, i.e., memory and a processor, located away from the aircraft, and a second computer having a computer having a computer and a processor, mounted on the aircraft, The first computer is, To provide a database describing the performance envelope of the aforementioned weapon, The process involves creating the coefficient characteristics of its performance envelope using a general algorithm, which includes a step of identifying the best candidate polynomial from several candidate polynomials, wherein the general algorithm has the form of a polynomial, and the variables of the polynomial are some or all of a group of firing condition parameters for a weapon or aircraft. Uploading the coefficients of the identified best candidate polynomial to the second computer. The second computer is configured to perform the following: A reconfigurator containing the same general algorithm selects coefficients for the general algorithm according to the conditions of the aircraft and the target, Using the selected coefficients, the reconstructor generates the feasibility representation. A system configured to perform the following, wherein the second computer is configured to select coefficients for the general algorithm according to the conditions of the aircraft and the target, according to the conditions of the aircraft and the target, by the reconfigurator which includes the same general algorithm, when the aircraft and the target are within the performance envelope of the weapon.

13. The system according to claim 12, further comprising a display for displaying the aforementioned feasibility indication.

14. An aircraft comprising the second computer according to claim 12.

15. A computer comprising a processor and memory configured to carry out the method described in any one of claims 1 to 11.

16. A computer program that, when executed by a computer having a processor and memory, comprises instructions causing the computer to perform the method according to any one of claims 1 to 11.

17. A non-transient computer-readable storage medium comprising an instruction, when executed by a computer having a processor and memory, causing the computer to perform the method according to any one of claims 1 to 11.

Images