Described are systems and methods for determining the gross weight of an aircraft. A flight regime is determined based on one or more inputs. A neural net is selected based on a flight regime. The neural net inputs may include derived values. A first estimate of the gross weight is produced by the s
Described are systems and methods for determining the gross weight of an aircraft. A flight regime is determined based on one or more inputs. A neural net is selected based on a flight regime. The neural net inputs may include derived values. A first estimate of the gross weight is produced by the selected neural net. The first estimate is used, along with other inputs, with a Kalman filter to produce a final gross weight estimate. The Kalman filter blends or fuses together its inputs to produce the final gross weight estimate.
대표청구항▼
What is claimed is: 1. A method for determining a weight of an aircraft comprising: determining a flight regime corresponding to a current flight state of the aircraft in accordance with one or more inputs characterizing said current flight state of the aircraft; selecting a neural net in accordanc
What is claimed is: 1. A method for determining a weight of an aircraft comprising: determining a flight regime corresponding to a current flight state of the aircraft in accordance with one or more inputs characterizing said current flight state of the aircraft; selecting a neural net in accordance with said flight regime, wherein said neural net is selected from a plurality of neural nets, each of said plurality of neural nets being trained to determine said weight of said aircraft for at least one flight regime based on a relationship between said weight and one or more neural net inputs, and wherein said neural net selected is trained to determine said weight of said aircraft in said flight regime; and determining said weight using said neural net. 2. The method of claim 1, wherein said neural net is trained offline prior to determining said weight of said aircraft. 3. The method of claim 2, wherein said determining said weight of said aircraft is performed during operation of said aircraft. 4. The method of claim 1, wherein said neural net is a feedforward neural net. 5. The method of claim 4, wherein said neural net includes a single hidden layer. 6. The method of claim 5, wherein said neural net has a same set of interconnections between each neuron in said hidden layer and an input layer, and a same set of interconnection between said each neuron and an output layer. 7. The method of claim 6, wherein each of said neurons in said hidden layer utilizes a same sigmoidal activation function. 8. The method of claim 7, wherein said neural net includes more than 20 neurons in said hidden layer. 9. The method of claim 1, wherein said weight is used as an input to another process. 10. The method of claim 1, wherein the flight regime is one of a plurality of flight regimes that are mutually exclusive from one another. 11. The method of claim 1, wherein the flight regime is manually selected. 12. The method of claim 1, wherein the flight regime is an effective flight regime including one or more actual flight regimes using the same set of one or more neural nets. 13. The method of claim 1, wherein one or more neural net inputs are used as inputs to said neural net selected, and the one or more neural net inputs include at least one derived parameter that is determined based on mathematical and physical relationships of measured data. 14. The method of claim 13, wherein the one or more neural net inputs are a first number of derived parameters determined using a second number of raw data values, the second number being greater than said first number. 15. The method of claim 13, wherein said one or more neural net inputs include at least one of the following: Corrected Vertical Acceleration (cNz) represented as: Where Nz is Vertical Acceleration; 횠 is Roll Attitude, wherein cos 횠 is not equal to zero; Torque Coefficient (Cq) represented as: Where Q is total torque (RPM); ρ is density (lb-sec2/ft4); A is the area of the main rotor disc (ft2); Ω is the rotation speed of the rotor (rad/s); R is the radius of the main rotor disc (ft); Nr is the main rotor speed (%); σ is the density ratio; and wherein ρ A(ΩR)2 is not equal to 0; Advance Ratio (μ) is represented as: Where KIAS is indicated airspeed in knots; and wherein ΩR is not equal to 0; Climb rate over tip speed (μc) is represented as: Where ROC is rate of climb (ft/min); and wherein ΩR is not equal to 0; Density Ratio (σ) is represented as: Where OAT is outside air temperature (째 C.); Hp is Barometric Altitude (ft). 16. The method of claim 15, wherein said neural net inputs include roll attitude and pitch attitude in accordance with the selected flight regime. 17. The method of claim 15, wherein one of said neural net inputs is a derived parameter based on at least one of roll attitude and pitch attitude in accordance with the selected flight regime. 18. The method of claim 1, wherein the neural net is included in a gross weight processor. 19. The method of claim 1, wherein the gross weight processor is included on the aircraft for which said weight is determined. 20. The method of claim 1, wherein the gross weight processor is included at a ground location and communicates with said aircraft. 21. The method of claim 1, wherein the one or more inputs include at least one of: a sensor measurement, manual input, data from a storage location. 22. The method of claim 1, further comprising: determining said flight regime as a hover flight regime in accordance with the following input parameters: landing flag, takeoff flag, weight on wheels, yaw rate, rate of climb, pitch attitude, roll attitude, drift velocity, ground speed, airspeed, and control reversal flag, wherein said landing flag indicates whether said aircraft is landing, said takeoff flag indicates whether said aircraft is in takeoff mode, and said control reversal flag indicates whether said aircraft is in a reversal mode. 23. The method of claim 22, wherein said landing flag indicates no landing, said takeoff flag indicates no takeoff, said weight on wheels indicates no weight on wheels, said control reversal flag indicates that said aircraft is not in reversal mode, said yaw rate has an approximate value within the inclusive range of:-2.5≦yaw rate 2.5 degrees/second, said pitch attitude is approximately 10 degrees, said rate of climb is approximately within the inclusive range of:-200≦rate of climb≦200 feet/minute, said roll attitude approximates a value within the inclusive range of:-6≦roll attitude≦3 degrees, said drift velocity approximates a value within the inclusive range of:-7≦drift velocity≦7 knots said ground speed approximates a value within the inclusive range of:-7≦ground speed≦7 knots, said airspeed is an approximate value less than or equal to 38 knots. 24. The method of claim 23, further comprising: determining that said aircraft is in a hover flight regime at a first point in time; and determining that said aircraft remains in said hover flight regime at a second later point in time if said airspeed at said second later point in time does not exceed 43 knots. 25. The method of claim 1, further comprising: determining said flight regime as a forward flight regime in accordance with the following input parameters: landing flag, takeoff flag, weight on wheels, yaw rate, rate of climb, pitch attitude, roll attitude, airspeed, control reversal flag, and sideslip, wherein said landing flag indicates whether said aircraft is landing, said takeoff flag indicates whether said aircraft is in takeoff mode, and said control reversal flag indicates whether said aircraft is in a reversal mode. 26. The method of claim 25, wherein said landing flag indicates no landing, said takeoff flag indicates no takeoff, said weight on wheels indicates no weight on wheels, said control reversal flag indicates that said aircraft is not in reversal mode, said yaw rate has an approximate value within the inclusive range of:-5≦yaw rate≦5 degrees/second, said pitch attitude is within the inclusive range of:-10≦pitch attitude≦10 degrees, said rate of climb is approximately within the inclusive range of:-500≦rate of climb≦500 feet/minute, said roll attitude approximates a value within the inclusive range of:-10≦roll attitude≦10 degrees, said side slip approximates a value within the inclusive range of:-0.05≦side slip≦0, said airspeed is an approximate value greater than 38 knots. 27. The method of claim 26, further comprising: determining that said aircraft is in a forward flight regime at a first point in time; and determining that said aircraft remains in said forward flight regime at a second later point in time if said airspeed at said second later point in time is greater than 33 knots. 28. The method of claim 1, further comprising: determining said flight regime as a turn flight regime in accordance with the following input parameters: landing flag, takeoff flag, weight on wheels, roll attitude, airspeed, and rate of climb, wherein said landing flag indicates whether said aircraft is landing and said takeoff flag indicates whether said aircraft is in takeoff mode. 29. The method of claim 28, wherein said landing flag indicates no landing, said takeoff flag indicates no takeoff, said weight on wheels indicates no weight on wheels, said rate of climb is approximately within the inclusive range of:-500≦rate of climb≦500 feet/minute, said roll attitude approximates a value within the inclusive range of:-10≦roll attitude≦10 degrees, said airspeed is an approximate value greater than 38 knots. 30. The method of claim 29, further comprising: determining that said aircraft is in a turn flight regime at a first point in time; and determining that said aircraft remains in said turn flight regime at a second later point in time unless at least one of the following is true: roll attitude is outside of the range-7,+13, and said airspeed is less than 36. 31. The method of claim 1, wherein said one or more inputs are scaled within a predetermined range. 32. The method of claim 1, further comprising: determining a sensitivity of said weight with respect to a parameter used in determining said weight. 33. The method of claim 32, wherein said sensitivity of said weight with respect to said parameter is determined in accordance with a partial derivative of said weight with respect to said parameter. 34. The method of claim 33, wherein said weight is determined using a neural network and represented as: where z is a vector of inputs, p is a number of neurons in the hidden layer, m is a number of inputs, W1i,j is a weight of the jth input to the ith neuron in the hidden layer, b1i is a bias added to the ith neuron, W2i is a weight of the ith neuron to the output neuron, b2 is a bias added to an output neuron, and γ is the tanh function, and wherein tanh is not equal to zero. 35. The method of claim 34, wherein, said neural network is a feedforward neural net with one hidden layer containing p sigmoidal neurons, and the sensitivity is represented as: where γ' is cosh-2, wherein tanh and cosh-2 are not equal to zero. 36. The method of claim 35, wherein said sensitivity with respect to an input vector z having said parameter that is a kth parameter, zk, is determined as a partial derivative of said weight with respect to the kth parameter evaluated in accordance with the input vector. 37. A method of determining a weight of an aircraft comprising: receiving one or more values related to the aircraft and used in determining the weight of the aircraft; and determining said weight at a point in time using a Kalman filter wherein said one or more values are used as inputs to said Kalman filter and said Kalman filter produces the weight as an output, a state as determined by said Kalman filter corresponding to state parameters including a weight of the aircraft at said point in time and fuel flow rate at said point in time, a measurement vector for said Kalman filter including a weight of the aircraft, a fuel flow rate, and a second weight of the aircraft determined in accordance with a flight regime of the aircraft characterizing a flight state of the aircraft, a covariance matrix used by said Kalman filter including an element associated with said second weight wherein a value of said element varies in accordance with whether said aircraft is determined to be in said flight regime when said determining step is performed. 38. The method of claim 37, wherein one or more measurements are input to said Kalman filter, and the method further comprising: selecting a function based on said flight regime; and determining, using said function, said value of said element of said covariance matrix associated with said second weight, said function using said one or more measurements to determine whether said aircraft is currently in said flight regime and to accordingly determine said value. 39. The method of claim 38, wherein said flight regime is the hover flight regime. 40. The method of claim 39, wherein said function determines said value of said element of said covariance matrix in accordance with body accelerations of said aircraft along x and z axes, roll attitude, pitch attitude, airspeed and altitude. 41. The method of claim 39, wherein said aircraft is a rotorcraft and said hover flight regime corresponds to said rotorcraft hovering. 42. The method of claim 38, wherein said flight regime is manually determined. 43. The method of claim 38, wherein said flight regime is determined in accordance with a predetermined mapping that maps one or more values to a particular flight regime, wherein a given set of one or more inputs values uniquely maps to a flight regime. 44. The method of claim 37, wherein said second weight of the aircraft is an estimate that is a predetermined value based on vehicle flight and performance data. 45. The method of claim 37, wherein said second weight of the aircraft is an estimate based on manually entered data representing a sum gross weight of said aircraft. 46. The method of claim 37, wherein said Kalman filter produces the weight as an output used as an input to another component. 47. A system for determining a weight of an aircraft comprising: a regime recognizer that determines a regime indicator in accordance with a portion of said one or more inputs, said regime indicator indicating a flight regime associated with a current flight state of the aircraft based on said portion of the one or more inputs; and a gross weight estimator that determines said weight of said aircraft, said gross weight estimator including at least one of: a Kalman filter, and one or more neural nets, and using at least one of said Kalman filter and a first of said one or more neural nets in determining said weight, each of said one or more neural nets being trained to determine said weight of the aircraft for at least one flight regime based on a relationship between said weight and one or more neural net inputs, a state as determined by said Kalman filter at a point in time corresponding to state parameters including the weight of the aircraft and fuel flow rate at said point in time. 48. The system of claim 47, wherein said system further comprises: an input processor that processes one or more inputs producing one or more processed inputs, said one or more inputs including at least one sensor measurement; and a portion of said one or more processed inputs are neural net inputs used by said one or more neural nets, and said gross weight estimator including: a neural net selector that selects a neural net in accordance with said regime indicator, said neural net selected being trained to determine said weight of said aircraft in said flight regime. 49. The system of claim 48, wherein said regime recognizer is included in said input processor. 50. The system of claim 47, wherein said gross weight estimator includes one or more neural nets whose output, when said one or more neural nets is selected in accordance with said flight regime indicator, is an input to said Kalman filter. 51. A method for determining an aircraft parameter comprising: determining a flight regime corresponding to a current flight state of an aircraft in accordance with one or more inputs characterizing said current flight state of the aircraft; selecting a neural net in accordance with said flight regime, wherein said neural net is selected from a plurality of neural nets, each of said plurality of neural nets being trained to determine said aircraft parameter when the aircraft is in at least one flight regime based on a relationship between said aircraft parameter and one or more neural net inputs, and wherein said neural net selected is trained to determine said aircraft parameter when the aircraft is in said flight regime; and determining said aircraft parameter using said neural net. 52. The method of claim 51, wherein said neural net uses at least one derived parameter determined from a relationship between one or more raw input values. 53. A system for determining an aircraft parameter comprising: a regime recognizer that determines a regime indicator in accordance with a portion of said one or more inputs, said regime indicator representing a flight regime of an aircraft being associated with a current flight state of the aircraft based on said portion of the one of more inputs; and an aircraft parameter generator that determines said aircraft parameter, said aircraft parameter generator including at least one of: a Kalman filter, and one or more neural nets, and using at least one of said Kalman filter and a first of said one or more neural nets in determining said aircraft parameter, each of said one or more neural nets being trained to determine said aircraft parameter when said aircraft is in at least one flight regime based on a relationship between said aircraft parameter and one or more neural net inputs, a state as determined by said Kalman filter including said aircraft parameter at a point in time. 54. A computer readable medium comprising code stored thereon for determining a weight of an aircraft, the computer readable medium comprising code that: determines a flight regime corresponding to a current flight state of the aircraft in accordance with one or more inputs characterizing said current flight state of the aircraft; selects a neural net in accordance with said flight regime, wherein said neural net is selected from a plurality of neural nets, each of said plurality of neural nets being trained to determine said weight of said aircraft for at least one flight regime based on a relationship between said weight and one or more neural net inputs, and wherein said neural net selected is trained to determine said weight of said aircraft in said flight regime; and determines said weight using said neural net. 55. The computer readable medium of claim 54, wherein said neural net is trained offline prior to determining said weight of said aircraft. 56. The computer readable medium of claim 55, wherein said code that determines said weight of said aircraft is executed during operation of said aircraft. 57. The computer readable medium of claim 54, wherein said neural net is a feedforward neural net. 58. The readable medium of claim 57, wherein said neural net includes a single hidden layer. 59. The computer readable medium of claim 58, wherein said neural net has a same set of interconnections between each neuron in said hidden layer and an input layer, and a same set of interconnection between said each neuron and an output layer. 60. The computer readable medium of claim 59, wherein each of said neurons in said hidden layer utilizes a same sigmoidal activation function. 61. The computer readable medium of claim 60, wherein said neural net includes more than 20 neurons in said hidden layer. 62. The computer readable medium of claim 54, wherein said weight is used as an input to another process. 63. The computer readable medium of claim 54, wherein the flight regime is one of a plurality of flight regimes that are mutually exclusive from one another. 64. The computer readable medium of claim 54, wherein the flight regime is manually selected. 65. The computer readable medium of claim 54, wherein the flight regime is an effective flight regime including one or more actual flight regimes using the same set of one or more neural nets. 66. The computer readable medium of claim 54, wherein one or more neural net inputs are used as inputs to said neural net selected, and the one or more neural net inputs include at least one derived parameter that is determined based on mathematical and physical relationships of measured data. 67. The computer readable medium of claim 66, wherein the one or more neural net inputs are a first number of derived parameters determined using a second number of raw data values, the second number being greater than said first number. 68. The computer readable medium of claim 66, wherein said one or more neural net inputs include at least one of the following: Corrected Vertical Acceleration (cNz) represented as: Where Nz is Vertical Acceleration; 횠 is Roll Attitude, wherein cos 횠 is not equal to zero; Torque Coefficient (Cq) represented as: Where Q is total torque (RPM); ρ is density (lb-sec2/ft4); A is the area of the main rotor disc (ft2); Ω is the rotation speed of the rotor (rad/s); R is the radius of the main rotor disc (ft); Nr is the main rotor speed (%); σ is the density ratio; and wherein ρ A(ΩR)2 not equal to 0; Advance Ratio (μ) is represented as: Where KIAS is indicated airspeed in knots; and wherein ΩR is not equal to 0; Climb rate over tip speed (μc) is represented as: Where ROC is rate of climb (ft/min); and wherein ΩR is not equal to 0; Density Ratio (σ) is represented as: Where OAT is outside air temperature (째 C.); Hp is Barometric Altitude (ft). 69. The computer readable medium of claim 68, wherein said neural net inputs include roll attitude and pitch attitude in accordance with the selected flight regime. 70. The computer readable medium of claim 68, wherein one of said neural net inputs is a derived parameter based on at least one of roll attitude and pitch attitude in accordance with the selected flight regime. 71. The computer readable medium of claim 54, wherein the neural net is included in a gross weight processor. 72. The computer readable medium of claim 54, wherein the gross weight processor is included on the aircraft for which said weight is determined. 73. The computer readable medium of claim 54, wherein the gross weight processor is included at a ground location and communicates with said aircraft. 74. The computer readable medium of claim 54, wherein the one or more inputs include at least one of: a sensor measurement, manual input, data from a storage location. 75. The computer readable medium of claim 54, further comprising code that: determines said flight regime as a hover flight regime in accordance with the following input parameters: landing flag, takeoff flag, weight on wheels, yaw rate, rate of climb, pitch attitude, roll attitude, drift velocity, ground speed, airspeed, and control reversal flag, wherein said landing flag indicates whether said aircraft is landing, said takeoff flag indicates whether said aircraft is in takeoff mode, and said control reversal flag indicates whether said aircraft is in a reversal mode. 76. The computer readable medium of claim 75, wherein said landing flag indicates no landing, said takeoff flag indicates no takeoff, said weight on wheels indicates no weight on wheels, said control reversal flag indicates that said aircraft is not in reversal mode, said yaw rate has an approximate value within the inclusive range of:-2.5≦yaw rate 2.5 degrees/second, said pitch attitude is approximately 10 degrees, said rate of climb is approximately within the inclusive range of:-200≦rate of climb≦200 feet/minute, said roll attitude approximates a value within the inclusive range of:-6≦roll attitude≦3 degrees, said drift velocity approximates a value within the inclusive range of:-7≦drift velocity≦7 knots said ground speed approximates a value within the inclusive range of:-7≦ground speed≦7 knots, said airspeed is an approximate value less than or equal to 38 knots. 77. The computer readable medium of claim 76, further comprising code that: determines said aircraft is in a hover flight regime at a first point in time; and determines said aircraft remains in said hover flight regime at a second later point in time if said airspeed at said second later point in time does not exceed 43 knots. 78. The computer readable medium of claim 54, further comprising: code that determines said flight regime as a forward flight regime in accordance with the following input parameters: landing flag, takeoff flag, weight on wheels, yaw rate, rate of climb, pitch attitude, roll attitude, airspeed, control reversal flag, and sideslip, wherein said landing flag indicates whether said aircraft is landing, said takeoff flag indicates whether said aircraft is in takeoff mode, and said control reversal flag indicates whether said aircraft is in a reversal mode. 79. The computer readable medium of claim 78, wherein said landing flag indicates no landing, said takeoff flag indicates no takeoff, said weight on wheels indicates no weight on wheels, said control reversal flag indicates that said aircraft is not in reversal mode, said yaw rate has an approximate value within the inclusive range of:-5≦yaw rate≦5 degrees/second, said pitch attitude is within the inclusive range of:-10≦pitch attitude≦10 degrees, said rate of climb is approximately within the inclusive range of:-500≦rate of climb≦500 feet/minute, said roll attitude approximates a value within the inclusive range of:-10≦roll attitude≦10 degrees, said side slip approximates a value within the inclusive range of:-0.05≦side slip≦0, said airspeed is an approximate value greater than 38 knots. 80. The computer readable medium of claim 79, further comprising code that: determines said aircraft is in a forward flight regime at a first point in time; and determines said aircraft remains in said forward flight regime at a second later point in time if said airspeed at said second later point in time is greater than 33 knots. 81. The computer readable medium of claim 54, further comprising code that: determines said flight regime as a turn flight regime in accordance with the following input parameters: landing flag, takeoff flag, weight on wheels, roll attitude, airspeed, and rate of climb, wherein said landing flag indicates whether said aircraft is landing and said takeoff flag indicates whether said aircraft is in takeoff mode. 82. The computer readable medium of claim 81, wherein said landing flag indicates no landing, said takeoff flag indicates no takeoff, said weight on wheels indicates no weight on wheels, said rate of climb is approximately within the inclusive range of:-500≦rate of climb≦500 feet/minute, said roll attitude approximates a value within the inclusive range of:-10≦roll attitude≦10 degrees, said airspeed is an approximate value greater than 38 knots. 83. The computer readable medium of claim 82, further comprising code that: determines said aircraft is in a turn flight regime at a first point in time; and determines said aircraft remains in said turn flight regime at a second later point in time unless at least one of the following is true: roll attitude is outside of the range-7,+13, and said airspeed is less than 36. 84. The computer readable medium of claim 54, wherein said one or more inputs are scaled within a predetermined range. 85. The computer readable medium of claim 54, further comprising code that: determines a sensitivity of said weight with respect to a parameter used in determining said weight. 86. The computer readable medium of claim 85, wherein said sensitivity of said weight with respect to said parameter is determined in accordance with a partial derivative of said weight with respect to said parameter. 87. The computer readable medium of claim 86, wherein said weight is determined using a neural network and represented as: where z is a vector of inputs, p is a number of neurons in the hidden layer, m is a number of inputs, W1i,j is a weight of the jth input to the ith neuron in the hidden layer, b1i is a bias added to the ith neuron, W2i is a weight of the ith neuron to the output neuron, b2 is a bias added to an output neuron, and γ is the tanh function and wherein tanh is not equal to zero. 88. The computer readable medium of claim 87, wherein, said neural network is a feedforward neural net with one hidden layer containing p sigmoidal neurons, and the sensitivity is represented as: where γ' is cosh-2, wherein tanh and cosh-2are not equal to zero. 89. The computer readable medium of claim 88, wherein said sensitivity with respect to an input vector z having said parameter that is a kth parameter, zk, is determined as a partial derivative of said weight with respect to the kth parameter evaluated in accordance with the input vector. 90. A computer readable medium comprising code stored thereon that determines a weight of an aircraft, the computer readable medium comprising code that: receives one or more values related to the aircraft and used in determining the weight of the aircraft; and determines said weight at a point in time using a Kalman filter wherein said one or more values are used as inputs to said Kalman filter and said Kalman filter produces the weight as an output, a state as determined by said Kalman filter corresponding to state parameters including a weight of the aircraft at said point in time and fuel flow rate at said point in time, a measurement vector for said Kalman filter including a weight of the aircraft, a fuel flow rate, and a second weight of the aircraft determined in accordance with a flight regime of the aircraft characterizing a flight state of the aircraft, a covariance matrix used by said Kalman filter including an element associated with said second weight wherein a value of said element varies in accordance with whether said aircraft is determined to be in said flight regime when said determines is performed. 91. The computer readable medium of claim 90, wherein one or more measurements are input to said Kalman filter, and the computer program product further comprising code that: selects a function based on said flight regime; and determines, using said function, said value of said element of said covariance matrix associated with said second weight, said function using one or more measurements to determine whether said aircraft is currently in said flight regime and accordingly determine said value. 92. The computer readable medium of claim 91, wherein said flight regime is the hover flight regime. 93. The computer readable medium of claim 92, wherein said function determines said value of said element of said covariance matrix in accordance with body accelerations of said aircraft along x and z axes, roll attitude, pitch attitude, airspeed and altitude. 94. The computer readable medium of claim 92, wherein said aircraft is a rotorcraft and said hover flight regime corresponds to said rotorcraft hovering. 95. The computer readable medium of claim 91, wherein said flight regime is manually determined. 96. The computer readable medium of claim 91, wherein said flight regime is determined in accordance with a predetermined mapping that maps one or more values to a particular flight regime, wherein a given set of one or more inputs values uniquely maps to a flight regime. 97. The computer readable medium of claim 90, wherein said second weight of the aircraft is an estimate that is a predetermined value based on vehicle flight and performance data. 98. The computer readable medium of claim 90, wherein said second weight of the aircraft is an estimate based on manually entered data representing a sum gross weight of said aircraft. 99. The computer readable medium of claim 90, wherein said Kalman filter produces the weight as an output used as an input to another component. 100. A computer readable medium comprising code stored thereon for determining an aircraft parameter, the computer readable medium comprising code that: determines a flight regime corresponding to a current flight state of an aircraft in accordance with one or more inputs characterizing said current flight state of the aircraft; selects a neural net in accordance with said flight regime, wherein said neural net is selected from a plurality of neural nets, each of said plurality of neural nets being trained to determine said aircraft parameter when the aircraft is in at least one flight regime based on a relationship between said aircraft parameter and one or more neural net inputs, and wherein said neural net selected is trained to determine said aircraft parameter when the aircraft is in said flight regime; and determines said aircraft parameter using said neural net. 101. The computer readable medium of claim 100, wherein said neural net uses at least one derived parameter determined from a relationship between one or more raw input values. 102. A method of determining a weight of an aircraft comprising: receiving one or more measurements characterizing a current flight state of the aircraft; determining a flight regime indicating the current flight state of the aircraft using at least a portion of said one or more measurements; selecting, in accordance with said flight regime, a neural net from a plurality of neural nets, each of said plurality of neural nets being trained to determine said weight of said aircraft for at least one flight regime based on a relationship between said weight and one or more neural net inputs, wherein said neural net selected is trained to determine said weight of said aircraft in said flight regime; determining said weight using said neural net; and using a Kalman filter to determine an updated weight of the aircraft, a state as determined by said Kalman filter corresponding to state parameters at a point in time including a weight of the aircraft and an engine fuel flow rate at said point in time, a measurement vector for said Kalman filter including said weight from said neural net. 103. The method of claim 102, wherein said aircraft is a rotorcraft. 104. The method of claim 102, wherein said Kalman filter uses a covariance matrix with an entry corresponding to said weight from said neural net, and wherein the method further comprising: determining a value for said entry using at least a second portion of said one or more measurements indicating a current flight state of the aircraft.
연구과제 타임라인
LOADING...
LOADING...
LOADING...
LOADING...
LOADING...
이 특허에 인용된 특허 (8)
Smith Joseph A. (Phoenix AZ) Ringnes Erik A. (Glendale AZ) Tarbutton Victor F. (Glendale AZ), Method of airplane performance estimation and prediction.
Jolly, Mark R.; Fuhrer, Zachary; Fogle, Robert E.; Catanzarite, David M., Aircraft rotary wing motion control and instrumented motion control fluid device.
James, Frances H.; Suh, John W.; Musgrove, Timothy A.; Ridge, Peter, Neighborhood guide for semantic search system and method to support local POI discovery.
※ AI-Helper는 부적절한 답변을 할 수 있습니다.