US20220222346A1 - Decentralized trust assessment - Google Patents
Decentralized trust assessment Download PDFInfo
- Publication number
- US20220222346A1 US20220222346A1 US17/584,248 US202217584248A US2022222346A1 US 20220222346 A1 US20220222346 A1 US 20220222346A1 US 202217584248 A US202217584248 A US 202217584248A US 2022222346 A1 US2022222346 A1 US 2022222346A1
- Authority
- US
- United States
- Prior art keywords
- trust
- neural network
- data stream
- trust module
- decentralized
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000013528 artificial neural network Methods 0.000 claims abstract description 66
- 238000012549 training Methods 0.000 claims description 24
- 238000012552 review Methods 0.000 claims description 15
- 238000000034 method Methods 0.000 claims description 9
- RZVHIXYEVGDQDX-UHFFFAOYSA-N 9,10-anthraquinone Chemical compound C1=CC=C2C(=O)C3=CC=CC=C3C(=O)C2=C1 RZVHIXYEVGDQDX-UHFFFAOYSA-N 0.000 description 10
- 238000001514 detection method Methods 0.000 description 5
- 238000003909 pattern recognition Methods 0.000 description 2
- 238000009825 accumulation Methods 0.000 description 1
- 230000004913 activation Effects 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 125000004122 cyclic group Chemical group 0.000 description 1
- 230000001351 cycling effect Effects 0.000 description 1
- 238000003066 decision tree Methods 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 230000003595 spectral effect Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F21/00—Security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
- G06F21/50—Monitoring users, programs or devices to maintain the integrity of platforms, e.g. of processors, firmware or operating systems
- G06F21/57—Certifying or maintaining trusted computer platforms, e.g. secure boots or power-downs, version controls, system software checks, secure updates or assessing vulnerabilities
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2221/00—Indexing scheme relating to security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
- G06F2221/03—Indexing scheme relating to G06F21/50, monitoring users, programs or devices to maintain the integrity of platforms
- G06F2221/031—Protect user input by software means
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2221/00—Indexing scheme relating to security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
- G06F2221/03—Indexing scheme relating to G06F21/50, monitoring users, programs or devices to maintain the integrity of platforms
- G06F2221/032—Protect output to user by software means
Definitions
- the trust assessment module is configured for accepting various input data streams and making quality determinations on those input data streams. Having only a trust assessment module limits the robustness of the system. For example, trust assessment modules look for data streams that have failed, are stuck at a value, or have reached a maximum or minimum. When the trust module has not been programmed to look for a specific condition, the trust module cannot detect it. Therefore, limitations exist in conventional trust assessment modules.
- FIG. 1 is an oblique view of a tiltrotor aircraft according to this disclosure.
- FIG. 2 is a side view of a rotorcraft according to this disclosure.
- FIGS. 3A and 3B are schematic views of a decentralized trust assessment system according to this disclosure.
- FIG. 4 is a schematic view of a decentralized trust assessment system according to this disclosure.
- FIG. 5 is a schematic view of a decentralized trust assessment system according to this disclosure.
- FIG. 6 is a schematic view of a decentralized trust assessment system according to this disclosure.
- FIG. 7 is a schematic view of a decentralized trust assessment system according to this disclosure.
- This disclosure teaches a system comprised of trust assessment modules in conjunction with neural networks.
- the improved system can identify when data streams meet a predetermined condition and when the data streams have formed a pattern worth being concerned over. Those data streams include inputs to the aircraft subsystem, outputs of the aircraft subsystem, and the state of the aircraft subsystem itself.
- the decentralized trust assessment system (DTAS) verifies that the aircraft subsystem is receiving good data and is not being spoofed by combining a trust module with a neural network.
- the system further verifies that the aircraft subsystem is generating good data.
- the system can override the faulty subsystem and provide a better quality output data stream.
- the trust module in combination with the neural network verifies the format, the authenticity, and the content of the inputs to the subsystem.
- the trust module in combination with the neural network verifies the subsystem behavior is appropriate.
- Neural networks do not require specific preprogramming to detect bad streams of data, however, their detection of bad streams of data is not absolute.
- Trust modules do require specific programming to detect bad streams of data, however, they cannot detect what they are not programmed for.
- Trust modules also can be a software component that is executed either within a processor of the subsystem or physically separate from it. Combining the two elements results in a superior airborne DTAS.
- FIG. 1 illustrates a tiltrotor aircraft 101 equipped with a decentralized trust assessment system (DTAS) 401 according to this disclosure.
- Aircraft 101 has a fuselage 103 with a cockpit 105 located in a forward portion of fuselage 103 .
- Wings 107 , 109 are attached to fuselage 103 , and an engine nacelle 111 , 113 is rotatably attached to the outer end of each wing 107 , 109 , respectively.
- Each nacelle 111 , 113 houses an engine (not shown), which is operably connected to a rotatable proprotor 115 , 117 .
- Each proprotor 115 , 117 comprises three blades 119 .
- Proprotors 115 , 117 rotate in opposite directions and comprise similar components, though components in proprotors 115 , 117 may be constructed and/or installed in a mirror, or reverse, manner from the opposite proprotor 115 , 117 .
- Aircraft 101 requires a plurality of flight control computers in conjunction with pilot inputs to fly the aircraft. Flight control computers rely on various sensors, such as pitot static airspeed, gyroscopes, global positioning sensors, accelerometers, thermocouples, etc. for providing conditional information to the flight control computers.
- An example is the flight control computer's use of airspeed to vary the speed of proprotors 115 , 117 .
- the airspeed system is verified by the DTAS 401 before the airspeed data is passed to the flight control computers. Therefore, the flight control computer can operate with a higher level of data confidence.
- FIG. 2 illustrates a rotorcraft 201 equipped with a decentralized trust assessment system (DTAS) 401 according to this disclosure.
- Rotorcraft 201 comprises a rotary system 203 carried by a fuselage 205 .
- One or more rotor blades 207 operably associated with rotor system 203 provide flight for rotorcraft 201 and are controlled with a plurality of control sticks within fuselage 205 feeding inputs into a flight control computer.
- a pilot can manipulate the cyclic stick 209 to change the pitch angle of rotor blades 207 , thus providing lateral and longitudinal flight direction, and/or manipulate pedals 211 for controlling yaw direction, furthermore the pilot can adjust the collective stick 213 to change the pitch angles of all of the rotor blades concurrently.
- the sticks and pedal movements are measured by potentiometer systems.
- the potentiometer systems feature a portion of the DTAS 401 and determine whether the data from the potentiometers is trusted. That trusted data is then provided to a flight control system having a portion of the DTAS 401 .
- FIG. 3A illustrates an untrusted training system 301 for a neural network of a decentralized trust assessment system (DTAS).
- Untrusted training system 301 is comprised of a subsystem 303 , a plurality of untrusted training sets 305 , and a trained neural network 307 .
- the plurality of untrusted training sets 305 is comprised of a summation of inputs to the subsystem 313 and outputs of the subsystem 315 .
- the plurality of untrusted training sets 305 are provided repetitively to the trained neural network 307 .
- the neural network reviews the plurality of untrusted training sets 305 learning to detect patterns in the plurality of untrusted training sets. For example, a swashplate actuator's control signal and a collective position signal can be inputs to the untrusted training system 301 .
- the subsystem might analyze the swashplate actuator's control signal and the collective position signal to check if the signals are hitting any maximums or minimums.
- the trained neural network 307 can analyze the signals to find a pattern where an amplitude of the collective position signal is decreasing while the swashplate actuator's control signal is increasing, thereby indicating an issue.
- FIG. 3B illustrates a trusted training system 331 for a neural network of a DTAS.
- Untrusted training system 331 is comprised of a subsystem 333 , a trust module 335 , a plurality of trusted training sets 337 , and a trusted trained neural network 339 .
- the plurality of untrusted training sets 337 is comprised of a summation of inputs to the subsystem 341 and outputs of the subsystem 343 .
- the plurality of trusted training sets 337 are provided repetitively to the trusted trained neural network 339 .
- the neural network reviews the plurality of trusted training sets 337 learning to detect patterns in the plurality of trusted training sets. For example, a trusted neural network can be developed for icing systems while the aircraft is completing icing testing.
- the trust module 335 adds additional confidence in the trusted trained neural network 339 because the trust module reviews incoming data streams into the local subsystem to validate the quality of the incoming data streams.
- local subsystem 333 is responsible for activation of an icing system to heat the wing upon accumulation of ice on the leading edges of the wings and the rotors.
- the trust module 335 is typically a preprocessor that ensures data and control signals are being processed within a set of bounds and within a set of expectations. Trust module 335 can be programmed to look at various thermocouples located across the wing.
- the trust module 335 utilizes elements such as neural network 339 , decision trees, artificial and machine intelligence methods, bounds checking, and other techniques rooted in software, firmware, and/or hardware to verify the incoming inputs and the provided inputs. Trust module 335 detects when any of those thermocouples are reporting an impossible or unlikely temperature, such as absolute zero, and in response the trust module can flag the thermocouple data as bad or questionable. Therefore, the local subsystem 333 will not use the failed thermocouple data. Trusted trained neural network 339 might detect that as thermocouples are failing, their outputs ramp down to absolute zero over a period of time. Together the trust module 335 and the trusted trained neural network 339 collectively work to detect failing sensors and failed sensors by the data they generate.
- elements such as neural network 339 , decision trees, artificial and machine intelligence methods, bounds checking, and other techniques rooted in software, firmware, and/or hardware to verify the incoming inputs and the provided inputs.
- Trust module 335 detects when any of those thermocouples are reporting an impossible or unlikely
- FIG. 4 illustrates a decentralized trust assessment system (DTAS) 401 .
- DTAS 401 is comprised of a subsystem 403 , a trust module 405 , a trained neural network 407 , a set of inputs 409 , and a set of outputs 411 .
- a trusted neural network Once a trusted neural network is trained as described above, it can be utilized in conjunction with a trust module to increase the reliability of various airborne systems on a rotorcraft or tiltrotor aircraft.
- the set of input data 409 is provided to both the trained neural network 407 and the trust module 405 for data quality reviews.
- the trust module 405 reviews the set of input data 409 for specific programmed elements such as data streams indicating maximums or minimums.
- the trained neural network 407 also reviews the set of input data 409 for pattern detection based upon the training of the trained neural network 407 .
- An output of the trained neural network 407 is provided to the trust module 405 to provide increased confidence in the trust module's assessment of a quality of the set of input data 409 .
- Local subsystem 403 operates based upon the trust module's 405 output and also provides data to the trust module 405 for consistency. Outputs of the trained neural network 407 , the trust module 405 , and the local subsystem 403 form the set of output data 411 .
- An example of the DTAS 401 uses accelerometers from a tilt-axis gearbox of a tiltrotor. Data streams from a plurality of accelerometers are fed to both the trained neural network 407 and the trust module 405 .
- the trust module 405 detects accelerometers that have failed or are providing data outside a predetermined max window.
- the trained neural network 407 spots when spectral patterns of the plurality of accelerometers are diverging away from each other, thereby indicating a failing gearbox.
- the outputs from the trust module 405 and the trained neural network 407 are provided to local subsystem 403 , for example, a gearbox monitoring system, to indicate a worn tilt-axis gearbox.
- FIG. 5 illustrates a decentralized trust assessment system (DTAS) 501 .
- DTAS 501 is comprised of a subsystem 503 , a trust module 505 , a trained neural network 507 located in the trust module 505 , a set of inputs 509 , and a set of outputs 511 .
- a trusted neural network Once a trusted neural network is trained as described above, it can be utilized inside trust module 505 to increase the reliability of various airborne systems on a rotorcraft or tiltrotor aircraft.
- the set of input data 509 is provided to the trust module 505 with the trained neural network 507 located inside the trust module 505 for data quality reviews.
- the trust module 505 reviews the set of input data 509 for specific programmed elements such as data streams indicating maximums or minimums.
- the trained neural network 507 also reviews the set of input data 509 for pattern detection based upon the training of the neural network.
- Local subsystem 503 operates based upon the trust module's 505 output and also provides data to the trust module 505 for consistency. All outputs of the trust module 505 and the local subsystem 503 form the set of outputs 511 .
- An example of the DTAS 501 uses for example, Aeronautical Radio, Incorporated (ARNIC) data from a flight control computer.
- Data streams from the flight control computer are fed to the trust module 505 .
- the trust module 505 detects bus channels that have failed or are providing data outside a predetermined max window.
- the trained neural network 507 located in the trust module 505 can spot when odd-numbered bus channels are cycling from min to max indicating a databus issue.
- the outputs from the trust module 505 are provided to local subsystem 403 , and indicate a bad or faulty ARNIC standard 429 data bus.
- FIG. 6 illustrates an alternative decentralized trust assessment system (DTAS) 601 .
- DTAS 601 is comprised of a subsystem 603 , a trust module 605 , a trusted trained neural network 607 located outside both the subsystem 603 and the trust module 605 , a set of inputs 609 , and a set of outputs 611 .
- a trusted neural network Once a trusted neural network is trained as described above it can be utilized to increase the reliability of various airborne systems on a rotorcraft or tiltrotor aircraft.
- the set of input data 609 is provided to the trust module 605 for data quality reviews.
- the trust module 605 reviews the set of input data 609 for specific programmed elements such as data streams indicating maximums or minimums.
- Local subsystem 603 operates based upon the trust module's 605 output and also provides data to the trust module 605 for consistency. All outputs of the trust module 605 and the local subsystem 603 form the set of outputs 611 .
- the set of outputs 611 are fed into the trained neural network 607 for pattern detection based upon the training of the neural network.
- FIG. 7 illustrates another alternative decentralized trust assessment system (DTAS) 701 .
- DTAS 701 is comprised of a subsystem 703 , a trust module 705 , a trusted trained neural network 707 located outside the subsystem 703 and the trust module 705 , a set of inputs 709 , and a set of outputs 711 .
- a trusted neural network 707 is trained as described above it can be utilized to increase the reliability of various airborne systems on a rotorcraft or tiltrotor aircraft.
- the set of input data 709 are provided to the trust module 705 for data quality reviews.
- the trust module 705 reviews the set of input data 709 for specific programmed elements such as data streams indicating maximums or minimums.
- Local subsystem 703 operates based upon the trust module's 705 output and also provides data to the trust module 705 for consistency. All outputs of the trust module 705 and the local subsystem 703 form the set of outputs 711 .
- the set of outputs 711 are fed into the trained neural network 707 for pattern detection based upon the training of the neural network. An output of the neural network is fed back into the set of inputs 709 and provides feedback to the local subsystem 703 .
- the decentralized trust assessment system described above increases the reliability of airborne systems located on aircraft and rotorcrafts.
- Neural networks alone increase the robustness of the aircraft by allowing pattern recognition to occur without specific programming to identify the pattern.
- Neural networks in conjunction with trust modules are combined to increase the robustness of the aircraft by allowing pattern recognition without specific programming and allowing the aircraft to detect bad data streams from failed systems and spoofing and allow the aircraft to deem sources trustworthy.
- R I lower limit
- R u upper limit
- any number falling within the range is specifically disclosed.
- R R I +k *(R u -R I ), wherein k is a variable ranging from 1 percent to 100 percent with a 1 percent increment, i.e., k is 1 percent, 2 percent, 3 percent, 4 percent, 5 percent, . . . 50 percent, 51 percent, 52 percent, . . . , 95 percent, 96 percent, 95 percent, 98 percent, 99 percent, or 100 percent.
- any numerical range defined by two R numbers as defined in the above is also specifically disclosed.
Abstract
Description
- Not applicable.
- Not applicable.
- Modern day aircraft require their avionics systems to be reliable because so much of the actual control of the aircraft is done by parts of the avionics system. Some conventional avionics systems utilize only a trust assessment module. The trust assessment module is configured for accepting various input data streams and making quality determinations on those input data streams. Having only a trust assessment module limits the robustness of the system. For example, trust assessment modules look for data streams that have failed, are stuck at a value, or have reached a maximum or minimum. When the trust module has not been programmed to look for a specific condition, the trust module cannot detect it. Therefore, limitations exist in conventional trust assessment modules.
-
FIG. 1 is an oblique view of a tiltrotor aircraft according to this disclosure. -
FIG. 2 is a side view of a rotorcraft according to this disclosure. -
FIGS. 3A and 3B are schematic views of a decentralized trust assessment system according to this disclosure. -
FIG. 4 is a schematic view of a decentralized trust assessment system according to this disclosure. -
FIG. 5 is a schematic view of a decentralized trust assessment system according to this disclosure. -
FIG. 6 is a schematic view of a decentralized trust assessment system according to this disclosure. -
FIG. 7 is a schematic view of a decentralized trust assessment system according to this disclosure. - In this disclosure, reference may be made to the spatial relationships between various components and to the spatial orientation of various aspects of components as the devices are depicted in the attached drawings. However, as will be recognized by those skilled in the art after a complete reading of this disclosure, the devices, members, apparatuses, etc. described herein may be positioned in any desired orientation. Thus, the use of terms such as “above,” “below,” “upper,” “lower,” or other like terms to describe a spatial relationship between various components or to describe the spatial orientation of aspects of such components should be understood to describe a relative relationship between the components or a spatial orientation of aspects of such components, respectively, as the device described herein may be oriented in any desired direction.
- This disclosure teaches a system comprised of trust assessment modules in conjunction with neural networks. The improved system can identify when data streams meet a predetermined condition and when the data streams have formed a pattern worth being concerned over. Those data streams include inputs to the aircraft subsystem, outputs of the aircraft subsystem, and the state of the aircraft subsystem itself. The decentralized trust assessment system (DTAS) verifies that the aircraft subsystem is receiving good data and is not being spoofed by combining a trust module with a neural network. The system further verifies that the aircraft subsystem is generating good data. The system can override the faulty subsystem and provide a better quality output data stream. The trust module in combination with the neural network verifies the format, the authenticity, and the content of the inputs to the subsystem. The trust module in combination with the neural network verifies the subsystem behavior is appropriate. Neural networks do not require specific preprogramming to detect bad streams of data, however, their detection of bad streams of data is not absolute. Trust modules do require specific programming to detect bad streams of data, however, they cannot detect what they are not programmed for. Trust modules also can be a software component that is executed either within a processor of the subsystem or physically separate from it. Combining the two elements results in a superior airborne DTAS.
-
FIG. 1 illustrates atiltrotor aircraft 101 equipped with a decentralized trust assessment system (DTAS) 401 according to this disclosure.Aircraft 101 has afuselage 103 with acockpit 105 located in a forward portion offuselage 103.Wings fuselage 103, and anengine nacelle wing nacelle rotatable proprotor proprotor blades 119.Proprotors proprotors opposite proprotor Aircraft 101 requires a plurality of flight control computers in conjunction with pilot inputs to fly the aircraft. Flight control computers rely on various sensors, such as pitot static airspeed, gyroscopes, global positioning sensors, accelerometers, thermocouples, etc. for providing conditional information to the flight control computers. An example is the flight control computer's use of airspeed to vary the speed ofproprotors -
FIG. 2 illustrates arotorcraft 201 equipped with a decentralized trust assessment system (DTAS) 401 according to this disclosure. Rotorcraft 201 comprises arotary system 203 carried by afuselage 205. One ormore rotor blades 207 operably associated withrotor system 203 provide flight forrotorcraft 201 and are controlled with a plurality of control sticks withinfuselage 205 feeding inputs into a flight control computer. For example, during flight a pilot can manipulate thecyclic stick 209 to change the pitch angle ofrotor blades 207, thus providing lateral and longitudinal flight direction, and/or manipulatepedals 211 for controlling yaw direction, furthermore the pilot can adjust thecollective stick 213 to change the pitch angles of all of the rotor blades concurrently. The sticks and pedal movements are measured by potentiometer systems. The potentiometer systems feature a portion of theDTAS 401 and determine whether the data from the potentiometers is trusted. That trusted data is then provided to a flight control system having a portion of theDTAS 401. -
FIG. 3A illustrates anuntrusted training system 301 for a neural network of a decentralized trust assessment system (DTAS).Untrusted training system 301 is comprised of asubsystem 303, a plurality ofuntrusted training sets 305, and a trainedneural network 307. - The plurality of
untrusted training sets 305 is comprised of a summation of inputs to thesubsystem 313 and outputs of thesubsystem 315. The plurality ofuntrusted training sets 305 are provided repetitively to the trainedneural network 307. The neural network reviews the plurality of untrusted training sets 305 learning to detect patterns in the plurality of untrusted training sets. For example, a swashplate actuator's control signal and a collective position signal can be inputs to theuntrusted training system 301. Conventionally the subsystem might analyze the swashplate actuator's control signal and the collective position signal to check if the signals are hitting any maximums or minimums. The trainedneural network 307 can analyze the signals to find a pattern where an amplitude of the collective position signal is decreasing while the swashplate actuator's control signal is increasing, thereby indicating an issue. -
FIG. 3B illustrates a trustedtraining system 331 for a neural network of a DTAS.Untrusted training system 331 is comprised of asubsystem 333, atrust module 335, a plurality of trusted training sets 337, and a trusted trainedneural network 339. - The plurality of untrusted training sets 337 is comprised of a summation of inputs to the
subsystem 341 and outputs of thesubsystem 343. The plurality of trusted training sets 337 are provided repetitively to the trusted trainedneural network 339. The neural network reviews the plurality of trusted training sets 337 learning to detect patterns in the plurality of trusted training sets. For example, a trusted neural network can be developed for icing systems while the aircraft is completing icing testing. - The
trust module 335 adds additional confidence in the trusted trainedneural network 339 because the trust module reviews incoming data streams into the local subsystem to validate the quality of the incoming data streams. For example,local subsystem 333 is responsible for activation of an icing system to heat the wing upon accumulation of ice on the leading edges of the wings and the rotors. Thetrust module 335 is typically a preprocessor that ensures data and control signals are being processed within a set of bounds and within a set of expectations.Trust module 335 can be programmed to look at various thermocouples located across the wing. Thetrust module 335 utilizes elements such asneural network 339, decision trees, artificial and machine intelligence methods, bounds checking, and other techniques rooted in software, firmware, and/or hardware to verify the incoming inputs and the provided inputs.Trust module 335 detects when any of those thermocouples are reporting an impossible or unlikely temperature, such as absolute zero, and in response the trust module can flag the thermocouple data as bad or questionable. Therefore, thelocal subsystem 333 will not use the failed thermocouple data. Trusted trainedneural network 339 might detect that as thermocouples are failing, their outputs ramp down to absolute zero over a period of time. Together thetrust module 335 and the trusted trainedneural network 339 collectively work to detect failing sensors and failed sensors by the data they generate. -
FIG. 4 illustrates a decentralized trust assessment system (DTAS) 401.DTAS 401 is comprised of asubsystem 403, atrust module 405, a trainedneural network 407, a set ofinputs 409, and a set ofoutputs 411. Once a trusted neural network is trained as described above, it can be utilized in conjunction with a trust module to increase the reliability of various airborne systems on a rotorcraft or tiltrotor aircraft. - The set of
input data 409 is provided to both the trainedneural network 407 and thetrust module 405 for data quality reviews. Thetrust module 405 reviews the set ofinput data 409 for specific programmed elements such as data streams indicating maximums or minimums. The trainedneural network 407 also reviews the set ofinput data 409 for pattern detection based upon the training of the trainedneural network 407. An output of the trainedneural network 407 is provided to thetrust module 405 to provide increased confidence in the trust module's assessment of a quality of the set ofinput data 409.Local subsystem 403 operates based upon the trust module's 405 output and also provides data to thetrust module 405 for consistency. Outputs of the trainedneural network 407, thetrust module 405, and thelocal subsystem 403 form the set ofoutput data 411. - An example of the
DTAS 401 uses accelerometers from a tilt-axis gearbox of a tiltrotor. Data streams from a plurality of accelerometers are fed to both the trainedneural network 407 and thetrust module 405. Thetrust module 405 detects accelerometers that have failed or are providing data outside a predetermined max window. The trainedneural network 407 spots when spectral patterns of the plurality of accelerometers are diverging away from each other, thereby indicating a failing gearbox. The outputs from thetrust module 405 and the trainedneural network 407 are provided tolocal subsystem 403, for example, a gearbox monitoring system, to indicate a worn tilt-axis gearbox. -
FIG. 5 illustrates a decentralized trust assessment system (DTAS) 501.DTAS 501 is comprised of asubsystem 503, atrust module 505, a trainedneural network 507 located in thetrust module 505, a set ofinputs 509, and a set ofoutputs 511. Once a trusted neural network is trained as described above, it can be utilized insidetrust module 505 to increase the reliability of various airborne systems on a rotorcraft or tiltrotor aircraft. - The set of
input data 509 is provided to thetrust module 505 with the trainedneural network 507 located inside thetrust module 505 for data quality reviews. Thetrust module 505 reviews the set ofinput data 509 for specific programmed elements such as data streams indicating maximums or minimums. The trainedneural network 507 also reviews the set ofinput data 509 for pattern detection based upon the training of the neural network.Local subsystem 503 operates based upon the trust module's 505 output and also provides data to thetrust module 505 for consistency. All outputs of thetrust module 505 and thelocal subsystem 503 form the set ofoutputs 511. - An example of the
DTAS 501 uses for example, Aeronautical Radio, Incorporated (ARNIC) data from a flight control computer. Data streams from the flight control computer are fed to thetrust module 505. Thetrust module 505 detects bus channels that have failed or are providing data outside a predetermined max window. The trainedneural network 507 located in thetrust module 505 can spot when odd-numbered bus channels are cycling from min to max indicating a databus issue. The outputs from thetrust module 505 are provided tolocal subsystem 403, and indicate a bad or faulty ARNIC standard 429 data bus. -
FIG. 6 illustrates an alternative decentralized trust assessment system (DTAS) 601.DTAS 601 is comprised of asubsystem 603, atrust module 605, a trusted trainedneural network 607 located outside both thesubsystem 603 and thetrust module 605, a set ofinputs 609, and a set ofoutputs 611. Once a trusted neural network is trained as described above it can be utilized to increase the reliability of various airborne systems on a rotorcraft or tiltrotor aircraft. - The set of
input data 609 is provided to thetrust module 605 for data quality reviews. Thetrust module 605 reviews the set ofinput data 609 for specific programmed elements such as data streams indicating maximums or minimums.Local subsystem 603 operates based upon the trust module's 605 output and also provides data to thetrust module 605 for consistency. All outputs of thetrust module 605 and thelocal subsystem 603 form the set ofoutputs 611. The set ofoutputs 611 are fed into the trainedneural network 607 for pattern detection based upon the training of the neural network. -
FIG. 7 illustrates another alternative decentralized trust assessment system (DTAS) 701.DTAS 701 is comprised of asubsystem 703, atrust module 705, a trusted trainedneural network 707 located outside thesubsystem 703 and thetrust module 705, a set ofinputs 709, and a set ofoutputs 711. Once a trustedneural network 707 is trained as described above it can be utilized to increase the reliability of various airborne systems on a rotorcraft or tiltrotor aircraft. - The set of
input data 709 are provided to thetrust module 705 for data quality reviews. Thetrust module 705 reviews the set ofinput data 709 for specific programmed elements such as data streams indicating maximums or minimums.Local subsystem 703 operates based upon the trust module's 705 output and also provides data to thetrust module 705 for consistency. All outputs of thetrust module 705 and thelocal subsystem 703 form the set ofoutputs 711. The set ofoutputs 711 are fed into the trainedneural network 707 for pattern detection based upon the training of the neural network. An output of the neural network is fed back into the set ofinputs 709 and provides feedback to thelocal subsystem 703. - It should be noted that the decentralized trust assessment system described above increases the reliability of airborne systems located on aircraft and rotorcrafts. Neural networks alone increase the robustness of the aircraft by allowing pattern recognition to occur without specific programming to identify the pattern. Neural networks in conjunction with trust modules are combined to increase the robustness of the aircraft by allowing pattern recognition without specific programming and allowing the aircraft to detect bad data streams from failed systems and spoofing and allow the aircraft to deem sources trustworthy.
- At least one embodiment is disclosed, and variations, combinations, and/or modifications of the embodiment(s) and/or features of the embodiment(s) made by a person having ordinary skill in the art are within the scope of this disclosure. Alternative embodiments that result from combining, integrating, and/or omitting features of the embodiment(s) are also within the scope of this disclosure. Where numerical ranges or limitations are expressly stated, such express ranges or limitations should be understood to include iterative ranges or limitations of like magnitude falling within the expressly stated ranges or limitations (e.g., from about 1 to about 10 includes, 2, 3, 4, etc.; greater than 0.10 includes 0.11, 0.12, 0.13, etc.). For example, whenever a numerical range with a lower limit, RI, and an upper limit, Ru, is disclosed, any number falling within the range is specifically disclosed. In particular, the following numbers within the range are specifically disclosed: R=RI+k *(Ru-RI), wherein k is a variable ranging from 1 percent to 100 percent with a 1 percent increment, i.e., k is 1 percent, 2 percent, 3 percent, 4 percent, 5 percent, . . . 50 percent, 51 percent, 52 percent, . . . , 95 percent, 96 percent, 95 percent, 98 percent, 99 percent, or 100 percent. Moreover, any numerical range defined by two R numbers as defined in the above is also specifically disclosed. Use of the term “optionally” with respect to any element of a claim means that the element is required, or alternatively, the element is not required, both alternatives being within the scope of the claim. Use of broader terms such as comprises, includes, and having should be understood to provide support for narrower terms such as consisting of, consisting essentially of, and comprised substantially of. Accordingly, the scope of protection is not limited by the description set out above but is defined by the claims that follow, that scope including all equivalents of the subject matter of the claims. Each and every claim is incorporated as further disclosure into the specification and the claims are embodiment(s) of the present invention. Also, the phrases “at least one of A, B, and C” and “A and/or B and/or C” should each be interpreted to include only A, only B, only C, or any combination of A, B, and C.
Claims (20)
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US17/584,248 US20220222346A1 (en) | 2018-10-01 | 2022-01-25 | Decentralized trust assessment |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US16/149,042 US11232207B2 (en) | 2018-10-01 | 2018-10-01 | Decentralized trust assessment |
US17/584,248 US20220222346A1 (en) | 2018-10-01 | 2022-01-25 | Decentralized trust assessment |
Related Parent Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US16/149,042 Continuation US11232207B2 (en) | 2018-10-01 | 2018-10-01 | Decentralized trust assessment |
Publications (1)
Publication Number | Publication Date |
---|---|
US20220222346A1 true US20220222346A1 (en) | 2022-07-14 |
Family
ID=69947549
Family Applications (2)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US16/149,042 Active 2040-04-15 US11232207B2 (en) | 2018-10-01 | 2018-10-01 | Decentralized trust assessment |
US17/584,248 Pending US20220222346A1 (en) | 2018-10-01 | 2022-01-25 | Decentralized trust assessment |
Family Applications Before (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US16/149,042 Active 2040-04-15 US11232207B2 (en) | 2018-10-01 | 2018-10-01 | Decentralized trust assessment |
Country Status (1)
Country | Link |
---|---|
US (2) | US11232207B2 (en) |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114154353B (en) * | 2022-02-07 | 2022-04-12 | 中国空气动力研究与发展中心低速空气动力研究所 | Optical fiber icing detection sensor data reconstruction method, system and storage medium |
Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6112140A (en) * | 1996-05-14 | 2000-08-29 | The Boeing Company | Flight management system providing for automatic control display unit backup utilizing structured data routing |
US6115656A (en) * | 1997-06-17 | 2000-09-05 | Mcdonnell Douglas Corporation | Fault recording and reporting method |
US20060112119A1 (en) * | 2004-11-22 | 2006-05-25 | The Boeing Company | System, method and computer program product for real-time event indentification and course of action interpretation |
US9730339B2 (en) * | 2012-07-25 | 2017-08-08 | Edmund David Burke | Common bus structure for avionics and satellites (CBSAS) |
US20170312614A1 (en) * | 2016-05-02 | 2017-11-02 | Bao Tran | Smart device |
US20190042867A1 (en) * | 2017-12-28 | 2019-02-07 | Yen-Kuang Chen | Multi-domain cascade convolutional neural network |
US20190049957A1 (en) * | 2018-03-30 | 2019-02-14 | Intel Corporation | Emotional adaptive driving policies for automated driving vehicles |
US20190171822A1 (en) * | 2017-12-01 | 2019-06-06 | KnowBe4, Inc. | Systems and methods for aida campaign controller intelligent records |
US20190171934A1 (en) * | 2017-12-01 | 2019-06-06 | KnowBe4, Inc. | Systems and methods for aida based analytics and reporting |
US20190173919A1 (en) * | 2017-12-01 | 2019-06-06 | KnowBe4, Inc. | Systems and methods for aida based second chance |
-
2018
- 2018-10-01 US US16/149,042 patent/US11232207B2/en active Active
-
2022
- 2022-01-25 US US17/584,248 patent/US20220222346A1/en active Pending
Patent Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6112140A (en) * | 1996-05-14 | 2000-08-29 | The Boeing Company | Flight management system providing for automatic control display unit backup utilizing structured data routing |
US6115656A (en) * | 1997-06-17 | 2000-09-05 | Mcdonnell Douglas Corporation | Fault recording and reporting method |
US20060112119A1 (en) * | 2004-11-22 | 2006-05-25 | The Boeing Company | System, method and computer program product for real-time event indentification and course of action interpretation |
US9730339B2 (en) * | 2012-07-25 | 2017-08-08 | Edmund David Burke | Common bus structure for avionics and satellites (CBSAS) |
US20170312614A1 (en) * | 2016-05-02 | 2017-11-02 | Bao Tran | Smart device |
US20190171822A1 (en) * | 2017-12-01 | 2019-06-06 | KnowBe4, Inc. | Systems and methods for aida campaign controller intelligent records |
US20190171934A1 (en) * | 2017-12-01 | 2019-06-06 | KnowBe4, Inc. | Systems and methods for aida based analytics and reporting |
US20190173919A1 (en) * | 2017-12-01 | 2019-06-06 | KnowBe4, Inc. | Systems and methods for aida based second chance |
US20190042867A1 (en) * | 2017-12-28 | 2019-02-07 | Yen-Kuang Chen | Multi-domain cascade convolutional neural network |
US20190049957A1 (en) * | 2018-03-30 | 2019-02-14 | Intel Corporation | Emotional adaptive driving policies for automated driving vehicles |
Also Published As
Publication number | Publication date |
---|---|
US11232207B2 (en) | 2022-01-25 |
US20200104502A1 (en) | 2020-04-02 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
EP2810873B1 (en) | System and method for assisting in rotor speed control | |
US9415881B2 (en) | Rotorcraft having an airspeed sensor located at the top of a tail fin of the rotorcraft | |
US20220222346A1 (en) | Decentralized trust assessment | |
US10479513B2 (en) | Helicopter rotor icing detection system and method | |
US10712752B2 (en) | High authority stability and control augmentation system | |
US11216011B2 (en) | Optimized trajectory to noise improvement with auto-takeoff | |
US10752341B2 (en) | Tip clearance harmonic estimation | |
US20160378113A1 (en) | Dynamic compliance monitoring of unmanned vehicles | |
EP2905224B1 (en) | Rotor state sensor system | |
US20160325830A1 (en) | Tail rotor failure recovery controller | |
US10279893B2 (en) | System and method for validating rotorcraft position coordinates | |
CN104900092A (en) | Method for determining a guidance law for obstacle avoidance by an aircraft, related computer program product, electronic system and aircraft | |
CN105334845B (en) | Mechanical dual-redundancy electrical four-redundancy angle-of-attack redundancy management method | |
US11361078B2 (en) | Hierarchical integrated trust assessment | |
US9481456B2 (en) | Relative acceleration blade position measurement | |
Battipede et al. | Neural networks for air data estimation: Test of neural network simulating real flight instruments | |
KR101153967B1 (en) | Detecting method of flight error and flying object for self-detecting error | |
US9682771B2 (en) | Controlling rotor blades of a swashplateless rotor | |
US9266608B2 (en) | Compensating for lead-lag in rotor system | |
Barndt et al. | Maneuver regime recognition development and verification for H-60 structural monitoring | |
KR0173292B1 (en) | Autopilot pitch control apparatus of pilotless helicopter | |
Malpica et al. | Helicopter non-unique trim strategies for blade-vortex interaction (BVI) noise reduction | |
Kang et al. | Control law modification according to flight test of small scaled tilt rotor uav | |
KR101605839B1 (en) | Integration test system for arinc429 communication interface | |
Mariano et al. | Application of pilot induced oscillations prediction criteria to rotorcraft |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
STPP | Information on status: patent application and granting procedure in general |
Free format text: DOCKETED NEW CASE - READY FOR EXAMINATION |
|
AS | Assignment |
Owner name: BELL HELICOPTER TEXTRON INC., TEXAS Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:MCNAIR, MICHAEL KEVIN;REEL/FRAME:060998/0173 Effective date: 20180927 |
|
AS | Assignment |
Owner name: BELL TEXTRON INC., TEXAS Free format text: CHANGE OF NAME;ASSIGNOR:BELL HELICOPTER TEXTRON INC.;REEL/FRAME:061553/0759 Effective date: 20191015 |
|
AS | Assignment |
Owner name: BELL TEXTRON RHODE ISLAND INC., RHODE ISLAND Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:BELL TEXTRON INC.;REEL/FRAME:061610/0985 Effective date: 20220411 Owner name: TEXTRON INNOVATIONS INC., RHODE ISLAND Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:BELL TEXTRON RHODE ISLAND INC.;REEL/FRAME:061611/0087 Effective date: 20220411 |
|
STPP | Information on status: patent application and granting procedure in general |
Free format text: NON FINAL ACTION MAILED |
|
STPP | Information on status: patent application and granting procedure in general |
Free format text: FINAL REJECTION MAILED |
|
STPP | Information on status: patent application and granting procedure in general |
Free format text: DOCKETED NEW CASE - READY FOR EXAMINATION |
|
STPP | Information on status: patent application and granting procedure in general |
Free format text: NON FINAL ACTION MAILED |
|
STPP | Information on status: patent application and granting procedure in general |
Free format text: RESPONSE TO NON-FINAL OFFICE ACTION ENTERED AND FORWARDED TO EXAMINER |
|
STPP | Information on status: patent application and granting procedure in general |
Free format text: FINAL REJECTION MAILED |